LADI
/
py-sqlalchemy
1
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
py-sqlalchemy/CHANGES_PRE_06

6344 lines
283 KiB
Plaintext
Raw Permalink Blame History

-*- coding: utf-8; fill-column: 68 -*-
===============
CHANGES PRE-0.6
===============
0.5.9
=====
- sql
- Fixed erroneous self_group() call in expression package.
[ticket:1661]
0.5.8
=====
- sql
- The copy() method on Column now supports uninitialized,
unnamed Column objects. This allows easy creation of
declarative helpers which place common columns on multiple
subclasses.
- Default generators like Sequence() translate correctly
across a copy() operation.
- Sequence() and other DefaultGenerator objects are accepted
as the value for the "default" and "onupdate" keyword
arguments of Column, in addition to being accepted
positionally.
- Fixed a column arithmetic bug that affected column
correspondence for cloned selectables which contain
free-standing column expressions. This bug is
generally only noticeable when exercising newer
ORM behavior only availble in 0.6 via [ticket:1568],
but is more correct at the SQL expression level
as well. [ticket:1617]
- postgresql
- The extract() function, which was slightly improved in
0.5.7, needed a lot more work to generate the correct
typecast (the typecasts appear to be necessary in PG's
EXTRACT quite a lot of the time). The typecast is
now generated using a rule dictionary based
on PG's documentation for date/time/interval arithmetic.
It also accepts text() constructs again, which was broken
in 0.5.7. [ticket:1647]
- firebird
- Recognize more errors as disconnections. [ticket:1646]
0.5.7
=====
- orm
- contains_eager() now works with the automatically
generated subquery that results when you say
"query(Parent).join(Parent.somejoinedsubclass)", i.e.
when Parent joins to a joined-table-inheritance subclass.
Previously contains_eager() would erroneously add the
subclass table to the query separately producing a
cartesian product. An example is in the ticket
description. [ticket:1543]
- query.options() now only propagate to loaded objects
for potential further sub-loads only for options where
such behavior is relevant, keeping
various unserializable options like those generated
by contains_eager() out of individual instance states.
[ticket:1553]
- Session.execute() now locates table- and
mapper-specific binds based on a passed
in expression which is an insert()/update()/delete()
construct. [ticket:1054]
- Session.merge() now properly overwrites a many-to-one or
uselist=False attribute to None if the attribute
is also None in the given object to be merged.
- Fixed a needless select which would occur when merging
transient objects that contained a null primary key
identifier. [ticket:1618]
- Mutable collection passed to the "extension" attribute
of relation(), column_property() etc. will not be mutated
or shared among multiple instrumentation calls, preventing
duplicate extensions, such as backref populators,
from being inserted into the list.
[ticket:1585]
- Fixed the call to get_committed_value() on CompositeProperty.
[ticket:1504]
- Fixed bug where Query would crash if a join() with no clear
"left" side were called when a non-mapped column entity
appeared in the columns list. [ticket:1602]
- Fixed bug whereby composite columns wouldn't load properly
when configured on a joined-table subclass, introduced in
version 0.5.6 as a result of the fix for [ticket:1480].
[ticket:1616] thx to Scott Torborg.
- The "use get" behavior of many-to-one relations, i.e. that a
lazy load will fallback to the possibly cached query.get()
value, now works across join conditions where the two compared
types are not exactly the same class, but share the same
"affinity" - i.e. Integer and SmallInteger. Also allows
combinations of reflected and non-reflected types to work
with 0.5 style type reflection, such as PGText/Text (note 0.6
reflects types as their generic versions). [ticket:1556]
- Fixed bug in query.update() when passing Cls.attribute
as keys in the value dict and using synchronize_session='expire'
('fetch' in 0.6). [ticket:1436]
- sql
- Fixed bug in two-phase transaction whereby commit() method
didn't set the full state which allows subsequent close()
call to succeed. [ticket:1603]
- Fixed the "numeric" paramstyle, which apparently is the
default paramstyle used by Informixdb.
- Repeat expressions in the columns clause of a select
are deduped based on the identity of each clause element,
not the actual string. This allows positional
elements to render correctly even if they all render
identically, such as "qmark" style bind parameters.
[ticket:1574]
- The cursor associated with connection pool connections
(i.e. _CursorFairy) now proxies `__iter__()` to the
underlying cursor correctly. [ticket:1632]
- types now support an "affinity comparison" operation, i.e.
that an Integer/SmallInteger are "compatible", or
a Text/String, PickleType/Binary, etc. Part of
[ticket:1556].
- Fixed bug preventing alias() of an alias() from being
cloned or adapted (occurs frequently in ORM operations).
[ticket:1641]
- sqlite
- sqlite dialect properly generates CREATE INDEX for a table
that is in an alternate schema. [ticket:1439]
- postgresql
- Added support for reflecting the DOUBLE PRECISION type,
via a new postgres.PGDoublePrecision object.
This is postgresql.DOUBLE_PRECISION in 0.6.
[ticket:1085]
- Added support for reflecting the INTERVAL YEAR TO MONTH
and INTERVAL DAY TO SECOND syntaxes of the INTERVAL
type. [ticket:460]
- Corrected the "has_sequence" query to take current schema,
or explicit sequence-stated schema, into account.
[ticket:1576]
- Fixed the behavior of extract() to apply operator
precedence rules to the "::" operator when applying
the "timestamp" cast - ensures proper parenthesization.
[ticket:1611]
- mssql
- Changed the name of TrustedConnection to
Trusted_Connection when constructing pyodbc connect
arguments [ticket:1561]
- oracle
- The "table_names" dialect function, used by MetaData
.reflect(), omits "index overflow tables", a system
table generated by Oracle when "index only tables"
with overflow are used. These tables aren't accessible
via SQL and can't be reflected. [ticket:1637]
- ext
- A column can be added to a joined-table declarative
superclass after the class has been constructed
(i.e. via class-level attribute assignment), and
the column will be propagated down to
subclasses. [ticket:1570] This is the reverse
situation as that of [ticket:1523], fixed in 0.5.6.
- Fixed a slight inaccuracy in the sharding example.
Comparing equivalence of columns in the ORM is best
accomplished using col1.shares_lineage(col2).
[ticket:1491]
- Removed unused `load()` method from ShardedQuery.
[ticket:1606]
0.5.6
=====
- orm
- Fixed bug whereby inheritance discriminator part of a
composite primary key would fail on updates.
Continuation of [ticket:1300].
- Fixed bug which disallowed one side of a many-to-many
bidirectional reference to declare itself as "viewonly"
[ticket:1507]
- Added an assertion that prevents a @validates function
or other AttributeExtension from loading an unloaded
collection such that internal state may be corrupted.
[ticket:1526]
- Fixed bug which prevented two entities from mutually
replacing each other's primary key values within a single
flush() for some orderings of operations. [ticket:1519]
- Fixed an obscure issue whereby a joined-table subclass
with a self-referential eager load on the base class
would populate the related object's "subclass" table with
data from the "subclass" table of the parent.
[ticket:1485]
- relations() now have greater ability to be "overridden",
meaning a subclass that explicitly specifies a relation()
overriding that of the parent class will be honored
during a flush. This is currently to support
many-to-many relations from concrete inheritance setups.
Outside of that use case, YMMV. [ticket:1477]
- Squeezed a few more unnecessary "lazy loads" out of
relation(). When a collection is mutated, many-to-one
backrefs on the other side will not fire off to load
the "old" value, unless "single_parent=True" is set.
A direct assignment of a many-to-one still loads
the "old" value in order to update backref collections
on that value, which may be present in the session
already, thus maintaining the 0.5 behavioral contract.
[ticket:1483]
- Fixed bug whereby a load/refresh of joined table
inheritance attributes which were based on
column_property() or similar would fail to evaluate.
[ticket:1480]
- Improved support for MapperProperty objects overriding
that of an inherited mapper for non-concrete
inheritance setups - attribute extensions won't randomly
collide with each other. [ticket:1488]
- UPDATE and DELETE do not support ORDER BY, LIMIT, OFFSET,
etc. in standard SQL. Query.update() and Query.delete()
now raise an exception if any of limit(), offset(),
order_by(), group_by(), or distinct() have been
called. [ticket:1487]
- Added AttributeExtension to sqlalchemy.orm.__all__
- Improved error message when query() is called with
a non-SQL /entity expression. [ticket:1476]
- Using False or 0 as a polymorphic discriminator now
works on the base class as well as a subclass.
[ticket:1440]
- Added enable_assertions(False) to Query which disables
the usual assertions for expected state - used
by Query subclasses to engineer custom state.
[ticket:1424]. See
http://www.sqlalchemy.org/trac/wiki/UsageRecipes/PreFilteredQuery
for an example.
- Fixed recursion issue which occured if a mapped object's
`__len__()` or `__nonzero__()` method resulted in state
changes. [ticket:1501]
- Fixed incorrect exception raise in
Weak/StrongIdentityMap.add()
[ticket:1506]
- Fixed the error message for "could not find a FROM clause"
in query.join() which would fail to issue correctly
if the query was against a pure SQL construct.
[ticket:1522]
- Fixed a somewhat hypothetical issue which would result
in the wrong primary key being calculated for a mapper
using the old polymorphic_union function - but this
is old stuff. [ticket:1486]
- sql
- Fixed column.copy() to copy defaults and onupdates.
[ticket:1373]
- Fixed a bug in extract() introduced in 0.5.4 whereby
the string "field" argument was getting treated as a
ClauseElement, causing various errors within more
complex SQL transformations.
- Unary expressions such as DISTINCT propagate their
type handling to result sets, allowing conversions like
unicode and such to take place. [ticket:1420]
- Fixed bug in Table and Column whereby passing empty
dict for "info" argument would raise an exception.
[ticket:1482]
- oracle
- Backported 0.6 fix for Oracle alias names not getting
truncated. [ticket:1309]
- ext
- The collection proxies produced by associationproxy are now
pickleable. A user-defined proxy_factory however
is still not pickleable unless it defines __getstate__
and __setstate__. [ticket:1446]
- Declarative will raise an informative exception if
__table_args__ is passed as a tuple with no dict argument.
Improved documentation. [ticket:1468]
- Table objects declared in the MetaData can now be used
in string expressions sent to primaryjoin/secondaryjoin/
secondary - the name is pulled from the MetaData of the
declarative base. [ticket:1527]
- A column can be added to a joined-table subclass after
the class has been constructed (i.e. via class-level
attribute assignment). The column is added to the underlying
Table as always, but now the mapper will rebuild its
"join" to include the new column, instead of raising
an error about "no such column, use column_property()
instead". [ticket:1523]
- test
- Added examples into the test suite so they get exercised
regularly and cleaned up a couple deprecation warnings.
0.5.5
=======
- general
- unit tests have been migrated from unittest to nose. See
README.unittests for information on how to run the tests.
[ticket:970]
- orm
- The "foreign_keys" argument of relation() will now propagate
automatically to the backref in the same way that primaryjoin
and secondaryjoin do. For the extremely rare use case where
the backref of a relation() has intentionally different
"foreign_keys" configured, both sides now need to be
configured explicity (if they do in fact require this setting,
see the next note...).
- ...the only known (and really, really rare) use case where a
different foreign_keys setting was used on the
forwards/backwards side, a composite foreign key that
partially points to its own columns, has been enhanced such
that the fk->itself aspect of the relation won't be used to
determine relation direction.
- Session.mapper is now *deprecated*.
Call session.add() if you'd like a free-standing object to be
part of your session. Otherwise, a DIY version of
Session.mapper is now documented at
http://www.sqlalchemy.org/trac/wiki/UsageRecipes/SessionAwareMapper
The method will remain deprecated throughout 0.6.
- Fixed Query being able to join() from individual columns of a
joined-table subclass entity, i.e. query(SubClass.foo,
SubcClass.bar).join(<anything>). In most cases, an error
"Could not find a FROM clause to join from" would be
raised. In a few others, the result would be returned in terms
of the base class rather than the subclass - so applications
which relied on this erroneous result need to be
adjusted. [ticket:1431]
- Fixed a bug involving contains_eager(), which would apply
itself to a secondary (i.e. lazy) load in a particular rare
case, producing cartesian products. improved the targeting of
query.options() on secondary loads overall [ticket:1461].
- Fixed bug introduced in 0.5.4 whereby Composite types fail
when default-holding columns are flushed.
- Fixed another 0.5.4 bug whereby mutable attributes
(i.e. PickleType) wouldn't be deserialized correctly when the
whole object was serialized. [ticket:1426]
- Fixed bug whereby session.is_modified() would raise an
exception if any synonyms were in use.
- Fixed potential memory leak whereby previously pickled objects
placed back in a session would not be fully garbage collected
unless the Session were explicitly closed out.
- Fixed bug whereby list-based attributes, like pickletype and
PGArray, failed to be merged() properly.
- Repaired non-working attributes.set_committed_value function.
- Trimmed the pickle format for InstanceState which should
further reduce the memory footprint of pickled instances. The
format should be backwards compatible with that of 0.5.4 and
previous.
- sqlalchemy.orm.join and sqlalchemy.orm.outerjoin are now
added to __all__ in sqlalchemy.orm.*. [ticket:1463]
- Fixed bug where Query exception raise would fail when
a too-short composite primary key value were passed to
get(). [ticket:1458]
- sql
- Removed an obscure feature of execute() (including connection,
engine, Session) whereby a bindparam() construct can be sent
as a key to the params dictionary. This usage is undocumented
and is at the core of an issue whereby the bindparam() object
created implicitly by a text() construct may have the same
hash value as a string placed in the params dictionary and may
result in an inappropriate match when computing the final bind
parameters. Internal checks for this condition would add
significant latency to the critical task of parameter
rendering, so the behavior is removed. This is a backwards
incompatible change for any application that may have been
using this feature, however the feature has never been
documented.
- engine/pool
- Implemented recreate() for StaticPool.
0.5.4p2
=======
- sql
- Repaired the printing of SQL exceptions which are not
based on parameters or are not executemany() style.
- postgresql
- Deprecated the hardcoded TIMESTAMP function, which when
used as func.TIMESTAMP(value) would render "TIMESTAMP value".
This breaks on some platforms as PostgreSQL doesn't allow
bind parameters to be used in this context. The hard-coded
uppercase is also inappropriate and there's lots of other
PG casts that we'd need to support. So instead, use
text constructs i.e. select(["timestamp '12/05/09'"]).
0.5.4p1
=======
- orm
- Fixed an attribute error introduced in 0.5.4 which would
occur when merge() was used with an incomplete object.
0.5.4
=====
- orm
- Significant performance enhancements regarding Sessions/flush()
in conjunction with large mapper graphs, large numbers of
objects:
- Removed all* O(N) scanning behavior from the flush() process,
i.e. operations that were scanning the full session,
including an extremely expensive one that was erroneously
assuming primary key values were changing when this
was not the case.
* one edge case remains which may invoke a full scan,
if an existing primary key attribute is modified
to a new value.
- The Session's "weak referencing" behavior is now *full* -
no strong references whatsoever are made to a mapped object
or related items/collections in its __dict__. Backrefs and
other cycles in objects no longer affect the Session's ability
to lose all references to unmodified objects. Objects with
pending changes still are maintained strongly until flush.
[ticket:1398]
The implementation also improves performance by moving
the "resurrection" process of garbage collected items
to only be relevant for mappings that map "mutable"
attributes (i.e. PickleType, composite attrs). This removes
overhead from the gc process and simplifies internal
behavior.
If a "mutable" attribute change is the sole change on an object
which is then dereferenced, the mapper will not have access to
other attribute state when the UPDATE is issued. This may present
itself differently to some MapperExtensions.
The change also affects the internal attribute API, but not
the AttributeExtension interface nor any of the publically
documented attribute functions.
- The unit of work no longer genererates a graph of "dependency"
processors for the full graph of mappers during flush(), instead
creating such processors only for those mappers which represent
objects with pending changes. This saves a tremendous number
of method calls in the context of a large interconnected
graph of mappers.
- Cached a wasteful "table sort" operation that previously
occured multiple times per flush, also removing significant
method call count from flush().
- Other redundant behaviors have been simplified in
mapper._save_obj().
- Modified query_cls on DynamicAttributeImpl to accept a full
mixin version of the AppenderQuery, which allows subclassing
the AppenderMixin.
- The "polymorphic discriminator" column may be part of a
primary key, and it will be populated with the correct
discriminator value. [ticket:1300]
- Fixed the evaluator not being able to evaluate IS NULL clauses.
- Fixed the "set collection" function on "dynamic" relations to
initiate events correctly. Previously a collection could only
be assigned to a pending parent instance, otherwise modified
events would not be fired correctly. Set collection is now
compatible with merge(), fixes [ticket:1352].
- Allowed pickling of PropertyOption objects constructed with
instrumented descriptors; previously, pickle errors would occur
when pickling an object which was loaded with a descriptor-based
option, such as query.options(eagerload(MyClass.foo)).
- Lazy loader will not use get() if the "lazy load" SQL clause
matches the clause used by get(), but contains some parameters
hardcoded. Previously the lazy strategy would fail with the
get(). Ideally get() would be used with the hardcoded
parameters but this would require further development.
[ticket:1357]
- MapperOptions and other state associated with query.options()
is no longer bundled within callables associated with each
lazy/deferred-loading attribute during a load.
The options are now associated with the instance's
state object just once when it's populated. This removes
the need in most cases for per-instance/attribute loader
objects, improving load speed and memory overhead for
individual instances. [ticket:1391]
- Fixed another location where autoflush was interfering
with session.merge(). autoflush is disabled completely
for the duration of merge() now. [ticket:1360]
- Fixed bug which prevented "mutable primary key" dependency
logic from functioning properly on a one-to-one
relation(). [ticket:1406]
- Fixed bug in relation(), introduced in 0.5.3,
whereby a self referential relation
from a base class to a joined-table subclass would
not configure correctly.
- Fixed obscure mapper compilation issue when inheriting
mappers are used which would result in un-initialized
attributes.
- Fixed documentation for session weak_identity_map -
the default value is True, indicating a weak
referencing map in use.
- Fixed a unit of work issue whereby the foreign
key attribute on an item contained within a collection
owned by an object being deleted would not be set to
None if the relation() was self-referential. [ticket:1376]
- Fixed Query.update() and Query.delete() failures with eagerloaded
relations. [ticket:1378]
- It is now an error to specify both columns of a binary primaryjoin
condition in the foreign_keys or remote_side collection. Whereas
previously it was just nonsensical, but would succeed in a
non-deterministic way.
- schema
- Added a quote_schema() method to the IdentifierPreparer class
so that dialects can override how schemas get handled. This
enables the MSSQL dialect to treat schemas as multipart
identifiers, such as 'database.owner'. [ticket: 594, 1341]
- sql
- Back-ported the "compiler" extension from SQLA 0.6. This
is a standardized interface which allows the creation of custom
ClauseElement subclasses and compilers. In particular it's
handy as an alternative to text() when you'd like to
build a construct that has database-specific compilations.
See the extension docs for details.
- Exception messages are truncated when the list of bound
parameters is larger than 10, preventing enormous
multi-page exceptions from filling up screens and logfiles
for large executemany() statements. [ticket:1413]
- ``sqlalchemy.extract()`` is now dialect sensitive and can
extract components of timestamps idiomatically across the
supported databases, including SQLite.
- Fixed __repr__() and other _get_colspec() methods on
ForeignKey constructed from __clause_element__() style
construct (i.e. declarative columns). [ticket:1353]
- mysql
- Reflecting a FOREIGN KEY construct will take into account
a dotted schema.tablename combination, if the foreign key
references a table in a remote schema. [ticket:1405]
- mssql
- Modified how savepoint logic works to prevent it from
stepping on non-savepoint oriented routines. Savepoint
support is still very experimental.
- Added in reserved words for MSSQL that covers version 2008
and all prior versions. [ticket:1310]
- Corrected problem with information schema not working with a
binary collation based database. Cleaned up information schema
since it is only used by mssql now. [ticket:1343]
- sqlite
- Corrected the SLBoolean type so that it properly treats only 1
as True. [ticket:1402]
- Corrected the float type so that it correctly maps to a
SLFloat type when being reflected. [ticket:1273]
- extensions
- Fixed adding of deferred or other column properties to a
declarative class. [ticket:1379]
0.5.3
=====
- orm
- The "objects" argument to session.flush() is deprecated.
State which represents the linkage between a parent and
child object does not support "flushed" status on
one side of the link and not the other, so supporting
this operation leads to misleading results.
[ticket:1315]
- Query now implements __clause_element__() which produces
its selectable, which means a Query instance can be accepted
in many SQL expressions, including col.in_(query),
union(query1, query2), select([foo]).select_from(query),
etc.
- Query.join() can now construct multiple FROM clauses, if
needed. Such as, query(A, B).join(A.x).join(B.y)
might say SELECT A.*, B.* FROM A JOIN X, B JOIN Y.
Eager loading can also tack its joins onto those
multiple FROM clauses. [ticket:1337]
- Fixed bug in dynamic_loader() where append/remove events
after construction time were not being propagated to the
UOW to pick up on flush(). [ticket:1347]
- Fixed bug where column_prefix wasn't being checked before
not mapping an attribute that already had class-level
name present.
- a session.expire() on a particular collection attribute
will clear any pending backref additions as well, so that
the next access correctly returns only what was present
in the database. Presents some degree of a workaround for
[ticket:1315], although we are considering removing the
flush([objects]) feature altogether.
- Session.scalar() now converts raw SQL strings to text()
the same way Session.execute() does and accepts same
alternative **kw args.
- improvements to the "determine direction" logic of
relation() such that the direction of tricky situations
like mapper(A.join(B)) -> relation-> mapper(B) can be
determined.
- When flushing partial sets of objects using session.flush([somelist]),
pending objects which remain pending after the operation won't
inadvertently be added as persistent. [ticket:1306]
- Added "post_configure_attribute" method to InstrumentationManager,
so that the "listen_for_events.py" example works again.
[ticket:1314]
- a forward and complementing backwards reference which are both
of the same direction, i.e. ONETOMANY or MANYTOONE,
is now detected, and an error message is raised.
Saves crazy CircularDependencyErrors later on.
- Fixed bugs in Query regarding simultaneous selection of
multiple joined-table inheritance entities with common base
classes:
- previously the adaption applied to "B" on
"A JOIN B" would be erroneously partially applied
to "A".
- comparisons on relations (i.e. A.related==someb)
were not getting adapted when they should.
- Other filterings, like
query(A).join(A.bs).filter(B.foo=='bar'), were erroneously
adapting "B.foo" as though it were an "A".
- Fixed adaptation of EXISTS clauses via any(), has(), etc.
in conjunction with an aliased object on the left and
of_type() on the right. [ticket:1325]
- Added an attribute helper method ``set_committed_value`` in
sqlalchemy.orm.attributes. Given an object, attribute name,
and value, will set the value on the object as part of its
"committed" state, i.e. state that is understood to have
been loaded from the database. Helps with the creation of
homegrown collection loaders and such.
- Query won't fail with weakref error when a non-mapper/class
instrumented descriptor is passed, raises
"Invalid column expession".
- Query.group_by() properly takes into account aliasing applied
to the FROM clause, such as with select_from(), using
with_polymorphic(), or using from_self().
- sql
- An alias() of a select() will convert to a "scalar subquery"
when used in an unambiguously scalar context, i.e. it's used
in a comparison operation. This applies to
the ORM when using query.subquery() as well.
- Fixed missing _label attribute on Function object, others
when used in a select() with use_labels (such as when used
in an ORM column_property()). [ticket:1302]
- anonymous alias names now truncate down to the max length
allowed by the dialect. More significant on DBs like
Oracle with very small character limits. [ticket:1309]
- the __selectable__() interface has been replaced entirely
by __clause_element__().
- The per-dialect cache used by TypeEngine to cache
dialect-specific types is now a WeakKeyDictionary.
This to prevent dialect objects from
being referenced forever for an application that
creates an arbitrarily large number of engines
or dialects. There is a small performance penalty
which will be resolved in 0.6. [ticket:1299]
- sqlite
- Fixed SQLite reflection methods so that non-present
cursor.description, which triggers an auto-cursor
close, will be detected so that no results doesn't
fail on recent versions of pysqlite which raise
an error when fetchone() called with no rows present.
- postgresql
- Index reflection won't fail when an index with
multiple expressions is encountered.
- Added PGUuid and PGBit types to
sqlalchemy.databases.postgres. [ticket:1327]
- Refection of unknown PG types won't crash when those
types are specified within a domain. [ticket:1327]
- mssql
- Preliminary support for pymssql 1.0.1
- Corrected issue on mssql where max_identifier_length was
not being respected.
- extensions
- Fixed a recursive pickling issue in serializer, triggered
by an EXISTS or other embedded FROM construct.
- Declarative locates the "inherits" class using a search
through __bases__, to skip over mixins that are local
to subclasses.
- Declarative figures out joined-table inheritance primary join
condition even if "inherits" mapper argument is given
explicitly.
- Declarative will properly interpret the "foreign_keys" argument
on a backref() if it's a string.
- Declarative will accept a table-bound column as a property
when used in conjunction with __table__, if the column is already
present in __table__. The column will be remapped to the given
key the same way as when added to the mapper() properties dict.
0.5.2
======
- orm
- Further refined 0.5.1's warning about delete-orphan cascade
placed on a many-to-many relation. First, the bad news:
the warning will apply to both many-to-many as well as
many-to-one relations. This is necessary since in both
cases, SQLA does not scan the full set of potential parents
when determining "orphan" status - for a persistent object
it only detects an in-python de-association event to establish
the object as an "orphan". Next, the good news: to support
one-to-one via a foreign key or assocation table, or to
support one-to-many via an association table, a new flag
single_parent=True may be set which indicates objects
linked to the relation are only meant to have a single parent.
The relation will raise an error if multiple parent-association
events occur within Python.
- Adjusted the attribute instrumentation change from 0.5.1 to
fully establish instrumentation for subclasses where the mapper
was created after the superclass had already been fully
instrumented. [ticket:1292]
- Fixed bug in delete-orphan cascade whereby two one-to-one
relations from two different parent classes to the same target
class would prematurely expunge the instance.
- Fixed an eager loading bug whereby self-referential eager
loading would prevent other eager loads, self referential or not,
from joining to the parent JOIN properly. Thanks to Alex K
for creating a great test case.
- session.expire() and related methods will not expire() unloaded
deferred attributes. This prevents them from being needlessly
loaded when the instance is refreshed.
- query.join()/outerjoin() will now properly join an aliased()
construct to the existing left side, even if query.from_self()
or query.select_from(someselectable) has been called.
[ticket:1293]
- sql
- Further fixes to the "percent signs and spaces in column/table
names" functionality. [ticket:1284]
- mssql
- Restored convert_unicode handling. Results were being passed
on through without conversion. [ticket:1291]
- Really fixing the decimal handling this time. [ticket:1282].
- Modified table reflection code to use only kwargs when
constructing tables. [Ticket:1289]
0.5.1
========
- orm
- Removed an internal join cache which could potentially leak
memory when issuing query.join() repeatedly to ad-hoc
selectables.
- The "clear()", "save()", "update()", "save_or_update()"
Session methods have been deprecated, replaced by
"expunge_all()" and "add()". "expunge_all()" has also
been added to ScopedSession.
- Modernized the "no mapped table" exception and added a more
explicit __table__/__tablename__ exception to declarative.
- Concrete inheriting mappers now instrument attributes which
are inherited from the superclass, but are not defined for
the concrete mapper itself, with an InstrumentedAttribute that
issues a descriptive error when accessed. [ticket:1237]
- Added a new `relation()` keyword `back_populates`. This
allows configuation of backreferences using explicit
relations. [ticket:781] This is required when creating
bidirectional relations between a hierarchy of concrete
mappers and another class. [ticket:1237]
- Test coverage added for `relation()` objects specified on
concrete mappers. [ticket:1237]
- Query.from_self() as well as query.subquery() both disable
the rendering of eager joins inside the subquery produced.
The "disable all eager joins" feature is available publically
via a new query.enable_eagerloads() generative. [ticket:1276]
- Added a rudimental series of set operations to Query that
receive Query objects as arguments, including union(),
union_all(), intersect(), except_(), insertsect_all(),
except_all(). See the API documentation for
Query.union() for examples.
- Fixed bug that prevented Query.join() and eagerloads from
attaching to a query that selected from a union or aliased union.
- A short documentation example added for bidirectional
relations specified on concrete mappers. [ticket:1237]
- Mappers now instrument class attributes upon construction
with the final InstrumentedAttribute object which remains
persistent. The `_CompileOnAttr`/`__getattribute__()`
methodology has been removed. The net effect is that
Column-based mapped class attributes can now be used fully
at the class level without invoking a mapper compilation
operation, greatly simplifying typical usage patterns
within declarative. [ticket:1269]
- ColumnProperty (and front-end helpers such as ``deferred``) no
longer ignores unknown **keyword arguments.
- Fixed a bug with the unitofwork's "row switch" mechanism,
i.e. the conversion of INSERT/DELETE into an UPDATE, when
combined with joined-table inheritance and an object
which contained no defined values for the child table where
an UPDATE with no SET clause would be rendered.
- Using delete-orphan on a many-to-many relation is deprecated.
This produces misleading or erroneous results since SQLA does
not retrieve the full list of "parents" for m2m. To get delete-orphan
behavior with an m2m table, use an explcit association class
so that the individual association row is treated as a parent.
[ticket:1281]
- delete-orphan cascade always requires delete cascade. Specifying
delete-orphan without delete now raises a deprecation warning.
[ticket:1281]
- sql
- Improved the methodology to handling percent signs in column
names from [ticket:1256]. Added more tests. MySQL and
PostgreSQL dialects still do not issue correct CREATE TABLE
statements for identifiers with percent signs in them.
- schema
- Index now accepts column-oriented InstrumentedAttributes
(i.e. column-based mapped class attributes) as column
arguments. [ticket:1214]
- Column with no name (as in declarative) won't raise a
NoneType error when it's string output is requsted
(such as in a stack trace).
- Fixed bug when overriding a Column with a ForeignKey
on a reflected table, where derived columns (i.e. the
"virtual" columns of a select, etc.) would inadvertently
call upon schema-level cleanup logic intended only
for the original column. [ticket:1278]
- declarative
- Can now specify Column objects on subclasses which have no
table of their own (i.e. use single table inheritance).
The columns will be appended to the base table, but only
mapped by the subclass.
- For both joined and single inheriting subclasses, the subclass
will only map those columns which are already mapped on the
superclass and those explicit on the subclass. Other
columns that are present on the `Table` will be excluded
from the mapping by default, which can be disabled
by passing a blank `exclude_properties` collection to the
`__mapper_args__`. This is so that single-inheriting
classes which define their own columns are the only classes
to map those columns. The effect is actually a more organized
mapping than you'd normally get with explicit `mapper()`
calls unless you set up the `exclude_properties` arguments
explicitly.
- It's an error to add new Column objects to a declarative class
that specified an existing table using __table__.
- mysql
- Added the missing keywords from MySQL 4.1 so they get escaped
properly.
- mssql
- Corrected handling of large decimal values with more robust
tests. Removed string manipulation on floats. [ticket:1280]
- Modified the do_begin handling in mssql to use the Cursor not
the Connection so it is DBAPI compatible.
- Corrected SAVEPOINT support on adodbapi by changing the
handling of savepoint_release, which is unsupported on mssql.
0.5.0
========
- new features
- general
- Documentation has been converted to Sphinx. In particular,
the generated API documentation has been constructed into a
full blown "API Reference" section which organizes editorial
documentation combined with generated docstrings. Cross
linking between sections and API docs are vastly improved, a
javascript-powered search feature is provided, and a full
index of all classes, functions and members is provided.
- setup.py now imports setuptools only optionally. If not
present, distutils is used. The new "pip" installer is
recommended over easy_install as it installs in a more
simplified way.
- added an extremely basic illustration of a PostGIS integration
to the examples folder.
- orm
- Query.with_polymorphic() now accepts a third argument
"discriminator" which will replace the value of
mapper.polymorphic_on for that query. Mappers themselves no
longer require polymorphic_on to be set, even if the mapper
has a polymorphic_identity. When not set, the mapper will
load non-polymorphically by default. Together, these two
features allow a non-polymorphic concrete inheritance setup to
use polymorphic loading on a per-query basis, since concrete
setups are prone to many issues when used polymorphically in
all cases.
- dynamic_loader accepts a query_class= to customize the Query
classes used for both the dynamic collection and the queries
built from it.
- query.order_by() accepts None which will remove any pending
order_by state from the query, as well as cancel out any
mapper/relation configured ordering. This is primarily useful
for overriding the ordering specified on a dynamic_loader().
[ticket:1079]
- sql
- RowProxy objects can be used in place of dictionary arguments
sent to connection.execute() and friends. [ticket:935]
- dialect
- Added a new description_encoding attribute on the dialect that
is used for encoding the column name when processing the
metadata. This usually defaults to utf-8.
- mssql
- Added in a new MSGenericBinary type. This maps to the Binary
type so it can implement the specialized behavior of treating
length specified types as fixed-width Binary types and
non-length types as an unbound variable length Binary type.
- Added in new types: MSVarBinary and MSImage. [ticket:1249]
- Added in the MSReal, MSNText, MSSmallDateTime, MSTime,
MSDateTimeOffset, and MSDateTime2 types
- sqlite
- Table reflection now stores the actual DefaultClause value for
the column. [ticket:1266]
- bugfixes, behavioral changes
- general
- orm
- Exceptions raised during compile_mappers() are now preserved
to provide "sticky behavior" - if a hasattr() call on a
pre-compiled mapped attribute triggers a failing compile and
suppresses the exception, subsequent compilation is blocked
and the exception will be reiterated on the next compile()
call. This issue occurs frequently when using declarative.
- property.of_type() is now recognized on a single-table
inheriting target, when used in the context of
prop.of_type(..).any()/has(), as well as
query.join(prop.of_type(...)).
- query.join() raises an error when the target of the join
doesn't match the property-based attribute - while it's
unlikely anyone is doing this, the SQLAlchemy author was
guilty of this particular loosey-goosey behavior.
- Fixed bug when using weak_instance_map=False where modified
events would not be intercepted for a flush(). [ticket:1272]
- Fixed some deep "column correspondence" issues which could
impact a Query made against a selectable containing multiple
versions of the same table, as well as unions and similar
which contained the same table columns in different column
positions at different levels. [ticket:1268]
- Custom comparator classes used in conjunction with
column_property(), relation() etc. can define new comparison
methods on the Comparator, which will become available via
__getattr__() on the InstrumentedAttribute. In the case of
synonym() or comparable_property(), attributes are resolved
first on the user-defined descriptor, then on the user-defined
comparator.
- Added ScopedSession.is_active accessor. [ticket:976]
- Can pass mapped attributes and column objects as keys to
query.update({}). [ticket:1262]
- Mapped attributes passed to the values() of an expression
level insert() or update() will use the keys of the mapped
columns, not that of the mapped attribute.
- Corrected problem with Query.delete() and Query.update() not
working properly with bind parameters. [ticket:1242]
- Query.select_from(), from_statement() ensure that the given
argument is a FromClause, or Text/Select/Union, respectively.
- Query() can be passed a "composite" attribute as a column
expression and it will be expanded. Somewhat related to
[ticket:1253].
- Query() is a little more robust when passed various column
expressions such as strings, clauselists, text() constructs
(which may mean it just raises an error more nicely).
- first() works as expected with Query.from_statement().
- Fixed bug introduced in 0.5rc4 involving eager loading not
functioning for properties which were added to a mapper
post-compile using add_property() or equivalent.
- Fixed bug where many-to-many relation() with viewonly=True
would not correctly reference the link between
secondary->remote.
- Duplicate items in a list-based collection will be maintained
when issuing INSERTs to a "secondary" table in a many-to-many
relation. Assuming the m2m table has a unique or primary key
constraint on it, this will raise the expected constraint
violation instead of silently dropping the duplicate
entries. Note that the old behavior remains for a one-to-many
relation since collection entries in that case don't result in
INSERT statements and SQLA doesn't manually police
collections. [ticket:1232]
- Query.add_column() can accept FromClause objects in the same
manner as session.query() can.
- Comparison of many-to-one relation to NULL is properly
converted to IS NOT NULL based on not_().
- Extra checks added to ensure explicit
primaryjoin/secondaryjoin are ClauseElement instances, to
prevent more confusing errors later on. [ticket:1087]
- Improved mapper() check for non-class classes. [ticket:1236]
- comparator_factory argument is now documented and supported by
all MapperProperty types, including column_property(),
relation(), backref(), and synonym() [ticket:5051].
- Changed the name of PropertyLoader to RelationProperty, to be
consistent with all the other names. PropertyLoader is still
present as a synonym.
- fixed "double iter()" call causing bus errors in shard API,
removed errant result.close() left over from the 0.4
version. [ticket:1099] [ticket:1228]
- made Session.merge cascades not trigger autoflush. Fixes
merged instances getting prematurely inserted with missing
values.
- Two fixes to help prevent out-of-band columns from being
rendered in polymorphic_union inheritance scenarios (which
then causes extra tables to be rendered in the FROM clause
causing cartesian products):
- improvements to "column adaption" for a->b->c inheritance
situations to better locate columns that are related to
one another via multiple levels of indirection, rather
than rendering the non-adapted column.
- the "polymorphic discriminator" column is only rendered
for the actual mapper being queried against. The column
won't be "pulled in" from a subclass or superclass mapper
since it's not needed.
- Fixed shard_id argument on ShardedSession.execute().
[ticket:1072]
- sql
- Columns can again contain percent signs within their
names. [ticket:1256]
- sqlalchemy.sql.expression.Function is now a public class. It
can be subclassed to provide user-defined SQL functions in an
imperative style, including with pre-established behaviors.
The postgis.py example illustrates one usage of this.
- PickleType now favors == comparison by default, if the
incoming object (such as a dict) implements __eq__(). If the
object does not implement __eq__() and mutable=True, a
deprecation warning is raised.
- Fixed the import weirdness in sqlalchemy.sql to not export
__names__ [ticket:1215].
- Using the same ForeignKey object repeatedly raises an error
instead of silently failing later. [ticket:1238]
- Added NotImplementedError for params() method on
Insert/Update/Delete constructs. These items currently don't
support this functionality, which also would be a little
misleading compared to values().
- Reflected foreign keys will properly locate their referenced
column, even if the column was given a "key" attribute
different from the reflected name. This is achieved via a new
flag on ForeignKey/ForeignKeyConstraint called "link_to_name",
if True means the given name is the referred-to column's name,
not its assigned key. [ticket:650]
- select() can accept a ClauseList as a column in the same way
as a Table or other selectable and the interior expressions
will be used as column elements. [ticket:1253]
- the "passive" flag on session.is_modified() is correctly
propagated to the attribute manager.
- union() and union_all() will not whack any order_by() that has
been applied to the select()s inside. If you union() a
select() with order_by() (presumably to support LIMIT/OFFSET),
you should also call self_group() on it to apply parenthesis.
- engine/pool
- Connection.invalidate() checks for closed status to avoid
attribute errors. [ticket:1246]
- NullPool supports reconnect on failure behavior.
[ticket:1094]
- Added a mutex for the initial pool creation when using
pool.manage(dbapi). This prevents a minor case of "dogpile"
behavior which would otherwise occur upon a heavy load
startup. [ticket:799]
- _execute_clauseelement() goes back to being a private method.
Subclassing Connection is not needed now that ConnectionProxy
is available.
- documentation
- Tickets [ticket:1200] [ticket:1149].
- Added note about create_session() defaults.
- Added section about metadata.reflect().
- Updated `TypeDecorator` section.
- Rewrote the "threadlocal" strategy section of the docs due to
recent confusion over this feature.
- Removed badly out of date 'polymorphic_fetch' and
'select_table' docs from inheritance, reworked the second half
of "joined table inheritance".
- Documented `comparator_factory` kwarg, added new doc section
"Custom Comparators".
- mssql
- Refactored the Date/Time types. The ``smalldatetime`` data
type no longer truncates to a date only, and will now be
mapped to the MSSmallDateTime type. [ticket:1254]
- Corrected an issue with Numerics to accept an int.
- Mapped ``char_length`` to the ``LEN()`` function.
- If an ``INSERT`` includes a subselect the ``INSERT`` is
converted from an ``INSERT INTO VALUES`` construct to a
``INSERT INTO SELECT`` construct.
- If the column is part of a ``primary_key`` it will be ``NOT
NULL`` since MSSQL doesn't allow ``NULL`` in primary_key
columns.
- ``MSBinary`` now returns a ``BINARY`` instead of an
``IMAGE``. This is a backwards incompatible change in that
``BINARY`` is a fixed length data type whereas ``IMAGE`` is a
variable length data type. [ticket:1249]
- ``get_default_schema_name`` is now reflected from the database
based on the user's default schema. This only works with MSSQL
2005 and later. [ticket:1258]
- Added collation support through the use of a new collation
argument. This is supported on the following types: char,
nchar, varchar, nvarchar, text, ntext. [ticket:1248]
- Changes to the connection string parameters favor DSN as the
default specification for pyodbc. See the mssql.py docstring
for detailed usage instructions.
- Added experimental support of savepoints. It currently does
not work fully with sessions.
- Support for three levels of column nullability: NULL, NOT
NULL, and the database's configured default. The default
Column configuration (nullable=True) will now generate NULL in
the DDL. Previously no specification was emitted and the
database default would take effect (usually NULL, but not
always). To explicitly request the database default,
configure columns with nullable=None and no specification will
be emitted in DDL. This is backwards incompatible
behavior. [ticket:1243]
- postgres
- "%" signs in text() constructs are automatically escaped to
"%%". Because of the backwards incompatible nature of this
change, a warning is emitted if '%%' is detected in the
string. [ticket:1267]
- Calling alias.execute() in conjunction with
server_side_cursors won't raise AttributeError.
- Added Index reflection support to PostgreSQL, using a great
patch we long neglected, submitted by Ken
Kuhlman. [ticket:714]
- oracle
- Adjusted the format of create_xid() to repair two-phase
commit. We now have field reports of Oracle two-phase commit
working properly with this change.
- Added OracleNVarchar type, produces NVARCHAR2, and also
subclasses Unicode so that convert_unicode=True by default.
NVARCHAR2 reflects into this type automatically so these
columns pass unicode on a reflected table with no explicit
convert_unicode=True flags. [ticket:1233]
- Fixed bug which was preventing out params of certain types
from being received; thanks a ton to huddlej at wwu.edu !
[ticket:1265]
- mysql
- "%" signs in text() constructs are automatically escaped to
"%%". Because of the backwards incompatible nature of this
change, a warning is emitted if '%%' is detected in the
string.
- Fixed bug in exception raise when FK columns not present
during reflection. [ticket:1241]
- Fixed bug involving reflection of a remote-schema table with a
foreign key ref to another table in that schema.
- associationproxy
- The association proxy properties are make themselves available
at the class level, e.g. MyClass.aproxy. Previously this
evaluated to None.
- declarative
- The full list of arguments accepted as string by backref()
includes 'primaryjoin', 'secondaryjoin', 'secondary',
'foreign_keys', 'remote_side', 'order_by'.
0.5.0rc4
========
- features
- orm
- Query.count() has been enhanced to do the "right thing" in a
wider variety of cases. It can now count multiple-entity
queries, as well as column-based queries. Note that this means
if you say query(A, B).count() without any joining criterion,
it's going to count the cartesian product of A*B. Any query
which is against column-based entities will automatically
issue "SELECT count(1) FROM (SELECT...)" so that the real
rowcount is returned, meaning a query such as
query(func.count(A.name)).count() will return a value of one,
since that query would return one row.
- Lots of performance tuning. A rough guesstimate over various
ORM operations places it 10% faster over 0.5.0rc3, 25-30% over
0.4.8.
- bugfixes and behavioral changes
- general:
- global "propigate"->"propagate" change.
- orm
- Adjustments to the enhanced garbage collection on
InstanceState to better guard against errors due to lost
state.
- Query.get() returns a more informative error message when
executed against multiple entities. [ticket:1220]
- Restored NotImplementedError on Cls.relation.in_()
[ticket:1140] [ticket:1221]
- Fixed PendingDeprecationWarning involving order_by parameter
on relation(). [ticket:1226]
- sql
- Removed the 'properties' attribute of the Connection object,
Connection.info should be used.
- Restored "active rowcount" fetch before ResultProxy autocloses
the cursor. This was removed in 0.5rc3.
- Rearranged the `load_dialect_impl()` method in `TypeDecorator`
such that it will take effect even if the user-defined
`TypeDecorator` uses another `TypeDecorator` as its impl.
- access
- Added support for Currency type.
- Functions were not return their result. [ticket:1017]
- Corrected problem with joins. Access only support LEFT OUTER
or INNER not just JOIN by itself. [ticket:1017]
- mssql
- Lots of cleanup and fixes to correct problems with limit and
offset.
- Correct situation where subqueries as part of a binary
expression need to be translated to use the IN and NOT IN
syntax.
- Fixed E Notation issue that prevented the ability to insert
decimal values less than 1E-6. [ticket:1216]
- Corrected problems with reflection when dealing with schemas,
particularly when those schemas are the default
schema. [ticket:1217]
- Corrected problem with casting a zero length item to a
varchar. It now correctly adjusts the CAST.
- ext
- Can now use a custom "inherit_condition" in __mapper_args__
when using declarative.
- fixed string-based "remote_side", "order_by" and others not
propagating correctly when used in backref().
0.5.0rc3
========
- features
- orm
- Added two new hooks to SessionExtension: after_bulk_delete()
and after_bulk_update(). after_bulk_delete() is called after
a bulk delete() operation on a query. after_bulk_update() is
called after a bulk update() operation on a query.
- sql
- SQL compiler optimizations and complexity reduction. The call
count for compiling a typical select() construct is 20% less
versus 0.5.0rc2.
- Dialects can now generate label names of adjustable
length. Pass in the argument "label_length=<value>" to
create_engine() to adjust how many characters max will be
present in dynamically generated column labels, i.e.
"somecolumn AS somelabel". Any value less than 6 will result
in a label of minimal size, consisting of an underscore and a
numeric counter. The compiler uses the value of
dialect.max_identifier_length as a default. [ticket:1211]
- ext
- Added a new extension sqlalchemy.ext.serializer. Provides
Serializer/Deserializer "classes" which mirror
Pickle/Unpickle, as well as dumps() and loads(). This
serializer implements an "external object" pickler which keeps
key context-sensitive objects, including engines, sessions,
metadata, Tables/Columns, and mappers, outside of the pickle
stream, and can later restore the pickle using any
engine/metadata/session provider. This is used not for
pickling regular object instances, which are pickleable
without any special logic, but for pickling expression objects
and full Query objects, such that all mapper/engine/session
dependencies can be restored at unpickle time.
- oracle
- Wrote a docstring for Oracle dialect. Apparently that Ohloh
"few source code comments" label is starting to sting :).
- Removed FIRST_ROWS() optimize flag when using LIMIT/OFFSET,
can be reenabled with optimize_limits=True create_engine()
flag. [ticket:536]
- bugfixes and behavioral changes
- orm
- "not equals" comparisons of simple many-to-one relation to an
instance will not drop into an EXISTS clause and will compare
foreign key columns instead.
- Removed not-really-working use cases of comparing a collection
to an iterable. Use contains() to test for collection
membership.
- Improved the behavior of aliased() objects such that they more
accurately adapt the expressions generated, which helps
particularly with self-referential comparisons. [ticket:1171]
- Fixed bug involving primaryjoin/secondaryjoin conditions
constructed from class-bound attributes (as often occurs when
using declarative), which later would be inappropriately
aliased by Query, particularly with the various EXISTS based
comparators.
- Fixed bug when using multiple query.join() with an
aliased-bound descriptor which would lose the left alias.
- Improved weakref identity map memory management to no longer
require mutexing, resurrects garbage collected instance on a
lazy basis for an InstanceState with pending changes.
- InstanceState object now removes circular references to itself
upon disposal to keep it outside of cyclic garbage collection.
- relation() won't hide unrelated ForeignKey errors inside of
the "please specify primaryjoin" message when determining join
condition.
- Fixed bug in Query involving order_by() in conjunction with
multiple aliases of the same class (will add tests in
[ticket:1218])
- When using Query.join() with an explicit clause for the ON
clause, the clause will be aliased in terms of the left side
of the join, allowing scenarios like query(Source).
from_self().join((Dest, Source.id==Dest.source_id)) to work
properly.
- polymorphic_union() function respects the "key" of each Column
if they differ from the column's name.
- Repaired support for "passive-deletes" on a many-to-one
relation() with "delete" cascade. [ticket:1183]
- Fixed bug in composite types which prevented a primary-key
composite type from being mutated [ticket:1213].
- Added more granularity to internal attribute access, such that
cascade and flush operations will not initialize unloaded
attributes and collections, leaving them intact for a
lazy-load later on. Backref events still initialize attrbutes
and collections for pending instances. [ticket:1202]
- sql
- Simplified the check for ResultProxy "autoclose without
results" to be based solely on presence of
cursor.description. All the regexp-based guessing about
statements returning rows has been removed [ticket:1212].
- Direct execution of a union() construct will properly set up
result-row processing. [ticket:1194]
- The internal notion of an "OID" or "ROWID" column has been
removed. It's basically not used by any dialect, and the
possibility of its usage with psycopg2's cursor.lastrowid is
basically gone now that INSERT..RETURNING is available.
- Removed "default_order_by()" method on all FromClause objects.
- Repaired the table.tometadata() method so that a passed-in
schema argument is propagated to ForeignKey constructs.
- Slightly changed behavior of IN operator for comparing to
empty collections. Now results in inequality comparison
against self. More portable, but breaks with stored procedures
that aren't pure functions.
- oracle
- Setting the auto_convert_lobs to False on create_engine() will
also instruct the OracleBinary type to return the cx_oracle
LOB object unchanged.
- mysql
- Fixed foreign key reflection in the edge case where a Table's
explicit schema= is the same as the schema (database) the
connection is attached to.
- No longer expects include_columns in table reflection to be
lower case.
- ext
- Fixed bug preventing declarative-bound "column" objects from
being used in column_mapped_collection(). [ticket:1174]
- misc
- util.flatten_iterator() func doesn't interpret strings with
__iter__() methods as iterators, such as in pypy
[ticket:1077].
0.5.0rc2
========
- orm
- Fixed bug involving read/write relation()s that contain
literal or other non-column expressions within their
primaryjoin condition equated to a foreign key column.
- "non-batch" mode in mapper(), a feature which allows mapper
extension methods to be called as each instance is
updated/inserted, now honors the insert order of the objects
given.
- Fixed RLock-related bug in mapper which could deadlock upon
reentrant mapper compile() calls, something that occurs when
using declarative constructs inside of ForeignKey objects.
- ScopedSession.query_property now accepts a query_cls factory,
overriding the session's configured query_cls.
- Fixed shared state bug interfering with ScopedSession.mapper's
ability to apply default __init__ implementations on object
subclasses.
- Fixed up slices on Query (i.e. query[x:y]) to work properly
for zero length slices, slices with None on either end.
[ticket:1177]
- Added an example illustrating Celko's "nested sets" as a
SQLA mapping.
- contains_eager() with an alias argument works even when
the alias is embedded in a SELECT, as when sent to the
Query via query.select_from().
- contains_eager() usage is now compatible with a Query that
also contains a regular eager load and limit/offset, in that
the columns are added to the Query-generated subquery.
[ticket:1180]
- session.execute() will execute a Sequence object passed to
it (regression from 0.4).
- Removed the "raiseerror" keyword argument from object_mapper()
and class_mapper(). These functions raise in all cases
if the given class/instance is not mapped.
- Fixed session.transaction.commit() on a autocommit=False
session not starting a new transaction.
- Some adjustments to Session.identity_map's weak referencing
behavior to reduce asynchronous GC side effects.
- Adjustment to Session's post-flush accounting of newly
"clean" objects to better protect against operating on
objects as they're asynchronously gc'ed. [ticket:1182]
- sql
- column.in_(someselect) can now be used as a columns-clause
expression without the subquery bleeding into the FROM clause
[ticket:1074]
- sqlite
- Overhauled SQLite date/time bind/result processing to use
regular expressions and format strings, rather than
strptime/strftime, to generically support pre-1900 dates,
dates with microseconds. [ticket:968]
- String's (and Unicode's, UnicodeText's, etc.) convert_unicode
logic disabled in the sqlite dialect, to adjust for pysqlite
2.5.0's new requirement that only Python unicode objects are
accepted;
http://itsystementwicklung.de/pipermail/list-pysqlite/2008-March/000018.html
- mysql
- Temporary tables are now reflectable.
- oracle
- Oracle will detect string-based statements which contain
comments at the front before a SELECT as SELECT statements.
[ticket:1187]
0.5.0rc1
========
- orm
- Query now has delete() and update(values) methods. This allows
to perform bulk deletes/updates with the Query object.
- The RowTuple object returned by Query(*cols) now features
keynames which prefer mapped attribute names over column keys,
column keys over column names, i.e. Query(Class.foo,
Class.bar) will have names "foo" and "bar" even if those are
not the names of the underlying Column objects. Direct Column
objects such as Query(table.c.col) will return the "key"
attribute of the Column.
- Added scalar() and value() methods to Query, each return a
single scalar value. scalar() takes no arguments and is
roughly equivalent to first()[0], value()
takes a single column expression and is roughly equivalent to
values(expr).next()[0].
- Improved the determination of the FROM clause when placing SQL
expressions in the query() list of entities. In particular
scalar subqueries should not "leak" their inner FROM objects
out into the enclosing query.
- Joins along a relation() from a mapped class to a mapped
subclass, where the mapped subclass is configured with single
table inheritance, will include an IN clause which limits the
subtypes of the joined class to those requested, within the ON
clause of the join. This takes effect for eager load joins as
well as query.join(). Note that in some scenarios the IN
clause will appear in the WHERE clause of the query as well
since this discrimination has multiple trigger points.
- AttributeExtension has been refined such that the event
is fired before the mutation actually occurs. Additionally,
the append() and set() methods must now return the given value,
which is used as the value to be used in the mutation operation.
This allows creation of validating AttributeListeners which
raise before the action actually occurs, and which can change
the given value into something else before its used.
- column_property(), composite_property(), and relation() now
accept a single or list of AttributeExtensions using the
"extension" keyword argument.
- query.order_by().get() silently drops the "ORDER BY" from
the query issued by GET but does not raise an exception.
- Added a Validator AttributeExtension, as well as a
@validates decorator which is used in a similar fashion
as @reconstructor, and marks a method as validating
one or more mapped attributes.
- class.someprop.in_() raises NotImplementedError pending the
implementation of "in_" for relation [ticket:1140]
- Fixed primary key update for many-to-many collections where
the collection had not been loaded yet [ticket:1127]
- Fixed bug whereby deferred() columns with a group in conjunction
with an otherwise unrelated synonym() would produce
an AttributeError during deferred load.
- The before_flush() hook on SessionExtension takes place before
the list of new/dirty/deleted is calculated for the final
time, allowing routines within before_flush() to further
change the state of the Session before the flush proceeds.
[ticket:1128]
- The "extension" argument to Session and others can now
optionally be a list, supporting events sent to multiple
SessionExtension instances. Session places SessionExtensions
in Session.extensions.
- Reentrant calls to flush() raise an error. This also serves
as a rudimentary, but not foolproof, check against concurrent
calls to Session.flush().
- Improved the behavior of query.join() when joining to
joined-table inheritance subclasses, using explicit join
criteria (i.e. not on a relation).
- @orm.attributes.reconstitute and
MapperExtension.reconstitute have been renamed to
@orm.reconstructor and MapperExtension.reconstruct_instance
- Fixed @reconstructor hook for subclasses which inherit from a
base class. [ticket:1129]
- The composite() property type now supports a
__set_composite_values__() method on the composite class which
is required if the class represents state using attribute
names other than the column's keynames; default-generated
values now get populated properly upon flush. Also,
composites with attributes set to None compare correctly.
[ticket:1132]
- The 3-tuple of iterables returned by attributes.get_history()
may now be a mix of lists and tuples. (Previously members
were always lists.)
- Fixed bug whereby changing a primary key attribute on an
entity where the attribute's previous value had been expired
would produce an error upon flush(). [ticket:1151]
- Fixed custom instrumentation bug whereby get_instance_dict()
was not called for newly constructed instances not loaded
by the ORM.
- Session.delete() adds the given object to the session if
not already present. This was a regression bug from 0.4.
[ticket:1150]
- The `echo_uow` flag on `Session` is deprecated, and unit-of-work
logging is now application-level only, not per-session level.
- Removed conflicting `contains()` operator from
`InstrumentedAttribute` which didn't accept `escape` kwaarg
[ticket:1153].
- declarative
- Fixed bug whereby mapper couldn't initialize if a composite
primary key referenced another table that was not defined
yet. [ticket:1161]
- Fixed exception throw which would occur when string-based
primaryjoin condition was used in conjunction with backref.
- schema
- Added "sorted_tables" accessor to MetaData, which returns
Table objects sorted in order of dependency as a list.
This deprecates the MetaData.table_iterator() method.
The "reverse=False" keyword argument has also been
removed from util.sort_tables(); use the Python
'reversed' function to reverse the results.
[ticket:1033]
- The 'length' argument to all Numeric types has been renamed
to 'scale'. 'length' is deprecated and is still accepted
with a warning.
- Dropped 0.3-compatibility for user defined types
(convert_result_value, convert_bind_param).
- sql
- Temporarily rolled back the "ORDER BY" enhancement from
[ticket:1068]. This feature is on hold pending further
development.
- The exists() construct won't "export" its contained list
of elements as FROM clauses, allowing them to be used more
effectively in the columns clause of a SELECT.
- and_() and or_() now generate a ColumnElement, allowing
boolean expressions as result columns, i.e.
select([and_(1, 0)]). [ticket:798]
- Bind params now subclass ColumnElement which allows them to be
selectable by orm.query (they already had most ColumnElement
semantics).
- Added select_from() method to exists() construct, which becomes
more and more compatible with a regular select().
- Added func.min(), func.max(), func.sum() as "generic functions",
which basically allows for their return type to be determined
automatically. Helps with dates on SQLite, decimal types,
others. [ticket:1160]
- added decimal.Decimal as an "auto-detect" type; bind parameters
and generic functions will set their type to Numeric when a
Decimal is used.
- mysql
- The 'length' argument to MSInteger, MSBigInteger, MSTinyInteger,
MSSmallInteger and MSYear has been renamed to 'display_width'.
- Added MSMediumInteger type [ticket:1146].
- the function func.utc_timestamp() compiles to UTC_TIMESTAMP, without
the parenthesis, which seem to get in the way when using in
conjunction with executemany().
- oracle
- limit/offset no longer uses ROW NUMBER OVER to limit rows,
and instead uses subqueries in conjunction with a special
Oracle optimization comment. Allows LIMIT/OFFSET to work
in conjunction with DISTINCT. [ticket:536]
- has_sequence() now takes the current "schema" argument into
account [ticket:1155]
- added BFILE to reflected type names [ticket:1121]
0.5beta3
========
0.5beta3 includes all bugfixes listed under release "0.4.7".
- orm
- The "entity_name" feature of SQLAlchemy mappers has been
removed. For rationale, see http://tinyurl.com/6nm2ne
- the "autoexpire" flag on Session, sessionmaker(), and
scoped_session() has been renamed to "expire_on_commit". It
does not affect the expiration behavior of rollback().
- fixed endless loop bug which could occur within a mapper's
deferred load of inherited attributes.
- a legacy-support flag "_enable_transaction_accounting" flag
added to Session which when False, disables all
transaction-level object accounting, including expire on
rollback, expire on commit, new/deleted list maintenance, and
autoflush on begin.
- The 'cascade' parameter to relation() accepts None as a value,
which is equivalent to no cascades.
- A critical fix to dynamic relations allows the "modified"
history to be properly cleared after a flush().
- user-defined @properties on a class are detected and left in
place during mapper initialization. This means that a
table-bound column of the same name will not be mapped at all
if a @property is in the way (and the column is not remapped
to a different name), nor will an instrumented attribute from
an inherited class be applied. The same rules apply for names
excluded using the include_properties/exclude_properties
collections.
- Added a new SessionExtension hook called after_attach(). This
is called at the point of attachment for objects via add(),
add_all(), delete(), and merge().
- A mapper which inherits from another, when inheriting the
columns of its inherited mapper, will use any reassigned
property names specified in that inheriting mapper.
Previously, if "Base" had reassigned "base_id" to the name
"id", "SubBase(Base)" would still get an attribute called
"base_id". This could be worked around by explicitly stating
the column in each submapper as well but this is fairly
unworkable and also impossible when using declarative
[ticket:1111].
- Fixed a series of potential race conditions in Session whereby
asynchronous GC could remove unmodified, no longer referenced
items from the session as they were present in a list of items
to be processed, typically during session.expunge_all() and
dependent methods.
- Some improvements to the _CompileOnAttr mechanism which should
reduce the probability of "Attribute x was not replaced during
compile" warnings. (this generally applies to SQLA hackers,
like Elixir devs).
- Fixed bug whereby the "unsaved, pending instance" FlushError
raised for a pending orphan would not take superclass mappers
into account when generating the list of relations responsible
for the error.
- sql
- func.count() with no arguments renders as COUNT(*), equivalent
to func.count(text('*')).
- simple label names in ORDER BY expressions render as
themselves, and not as a re-statement of their corresponding
expression. This feature is currently enabled only for
SQLite, MySQL, and PostgreSQL. It can be enabled on other
dialects as each is shown to support this
behavior. [ticket:1068]
- ext
- Class-bound attributes sent as arguments to relation()'s
remote_side and foreign_keys parameters are now accepted,
allowing them to be used with declarative. Additionally fixed
bugs involving order_by being specified as a class-bound
attribute in conjunction with eager loading.
- declarative initialization of Columns adjusted so that
non-renamed columns initialize in the same way as a non
declarative mapper. This allows an inheriting mapper to set
up its same-named "id" columns in particular such that the
parent "id" column is favored over the child column, reducing
database round trips when this value is requested.
- mysql
- Quoting of MSEnum values for use in CREATE TABLE is now
optional & will be quoted on demand as required. (Quoting was
always optional for use with existing tables.) [ticket:1110]
0.5beta2
========
0.5beta2 includes some of the bugfixes listed under release "0.4.7".
- orm
- In addition to expired attributes, deferred attributes also
load if their data is present in the result set. [ticket:870]
- session.refresh() raises an informative error message if the
list of attributes does not include any column-based
attributes.
- query() raises an informative error message if no columns or
mappers are specified.
- lazy loaders now trigger autoflush before proceeding. This
allows expire() of a collection or scalar relation to function
properly in the context of autoflush.
- column_property() attributes which represent SQL expressions
or columns that are not present in the mapped tables (such as
those from views) are automatically expired after an INSERT or
UPDATE, assuming they have not been locally modified, so that
they are refreshed with the most recent data upon access.
[ticket:887]
- Fixed explicit, self-referential joins between two
joined-table inheritance mappers when using query.join(cls,
aliased=True). [ticket:1082]
- Fixed query.join() when used in conjunction with a
columns-only clause and an SQL-expression ON clause in the
join.
- The "allow_column_override" flag from mapper() has been
removed. This flag is virtually always misunderstood. Its
specific functionality is available via the
include_properties/exclude_properties mapper arguments.
- Repaired `__str__()` method on Query. [ticket:1066]
- Session.bind gets used as a default even when table/mapper
specific binds are defined.
- schema
- Added prefixes option to `Table` that accepts a list of
strings to insert after CREATE in the CREATE TABLE statement.
[ticket:1075]
- Unicode, UnicodeText types now set "assert_unicode" and
"convert_unicode" by default, but accept overriding
**kwargs for these values.
- sql
- Added new match() operator that performs a full-text search.
Supported on PostgreSQL, SQLite, MySQL, MS-SQL, and Oracle
backends.
- sqlite
- Modified SQLite's representation of "microseconds" to match
the output of str(somedatetime), i.e. in that the microseconds
are represented as fractional seconds in string format. This
makes SQLA's SQLite date type compatible with datetimes that
were saved directly using Pysqlite (which just calls str()).
Note that this is incompatible with the existing microseconds
values in a SQLA 0.4 generated SQLite database file.
To get the old behavior globally:
from sqlalchemy.databases.sqlite import DateTimeMixin
DateTimeMixin.__legacy_microseconds__ = True
To get the behavior on individual DateTime types:
t = sqlite.SLDateTime()
t.__legacy_microseconds__ = True
Then use "t" as the type on the Column.
[ticket:1090]
- SQLite Date, DateTime, and Time types only accept Python
datetime objects now, not strings. If you'd like to format
dates as strings yourself with SQLite, use a String type. If
you'd like them to return datetime objects anyway despite
their accepting strings as input, make a TypeDecorator around
String - SQLA doesn't encourage this pattern.
- extensions
- Declarative supports a __table_args__ class variable, which is
either a dictionary, or tuple of the form (arg1, arg2, ...,
{kwarg1:value, ...}) which contains positional + kw arguments
to be passed to the Table constructor. [ticket:1096]
0.5beta1
========
An ongoing document describing the changes from 0.4 to 0.5 is at:
http://www.sqlalchemy.org/trac/wiki/05Migration
For this section, the list below primarily lists changes within the
attribute instrumentation API and is only a small fraction of the
full changes between 0.4/0.5.
-orm
- The "__init__" trigger/decorator added by mapper now attempts
to exactly mirror the argument signature of the original
__init__. The pass-through for '_sa_session' is no longer
implicit- you must allow for this keyword argument in your
constructor.
- ClassState is renamed to ClassManager.
- Classes may supply their own InstrumentationManager by
providing a __sa_instrumentation_manager__ property.
- Custom instrumentation may use any mechanism to associate a
ClassManager with a class and an InstanceState with an
instance. Attributes on those objects are still the default
association mechanism used by SQLAlchemy's native
instrumentation.
- Moved entity_name, _sa_session_id, and _instance_key from the
instance object to the instance state. These values are still
available in the old way, which is now deprecated, using
descriptors attached to the class. A deprecation warning will
be issued when accessed.
- The _prepare_instrumentation alias for prepare_instrumentation
has been removed.
- sqlalchemy.exceptions has been renamed to sqlalchemy.exc. The
module may be imported under either name.
- ORM-related exceptions are now defined in sqlalchemy.orm.exc.
ConcurrentModificationError, FlushError, and
UnmappedColumnError compatibility aliases are installed in
sqlalchemy.exc during the import of sqlalchemy.orm.
- sqlalchemy.logging has been renamed to sqlalchemy.log.
- The transitional sqlalchemy.log.SADeprecationWarning alias for
the warning's definition in sqlalchemy.exc has been removed.
- exc.AssertionError has been removed and usage replaced with
Python's built-in AssertionError.
- The behavior of MapperExtensions attached to multiple,
entity_name= primary mappers for a single class has been
altered. The first mapper() defined for a class is the only
mapper eligible for the MapperExtension 'instrument_class',
'init_instance' and 'init_failed' events. This is backwards
incompatible; previously the extensions of last mapper defined
would receive these events.
- firebird
- Added support for returning values from inserts (2.0+ only),
updates and deletes (2.1+ only).
0.4.9 (unreleased)
==================
- general:
- global "propigate"->"propagate" change.
- orm
- polymorphic_union() function respects the "key" of each
Column if they differ from the column's name.
- Fixed 0.4-only bug preventing composite columns
from working properly with inheriting mappers
[ticket:1199]
- Fixed RLock-related bug in mapper which could deadlock upon
reentrant mapper compile() calls, something that occurs when
using declarative constructs inside of ForeignKey objects.
Ported from 0.5.
- Fixed bug in composite types which prevented a primary-key
composite type from being mutated [ticket:1213].
- Added ScopedSession.is_active accessor. [ticket:976]
- Class-bound accessor can be used as the argument to
relation() order_by. [ticket:939]
- Fixed shard_id argument on ShardedSession.execute().
[ticket:1072]
- sql
- Connection.invalidate() checks for closed status
to avoid attribute errors. [ticket:1246]
- NullPool supports reconnect on failure behavior.
[ticket:1094]
- The per-dialect cache used by TypeEngine to cache
dialect-specific types is now a WeakKeyDictionary.
This to prevent dialect objects from
being referenced forever for an application that
creates an arbitrarily large number of engines
or dialects. There is a small performance penalty
which will be resolved in 0.6. [ticket:1299]
-sqlite
- Fixed SQLite reflection methods so that non-present
cursor.description, which triggers an auto-cursor
close, will be detected so that no results doesn't
fail on recent versions of pysqlite which raise
an error when fetchone() called with no rows present.
- postgres
- Added Index reflection support to Postgres, using a
great patch we long neglected, submitted by
Ken Kuhlman. [ticket:714]
- mysql
- Fixed bug in exception raise when FK columns not present
during reflection. [ticket:1241]
- oracle
- Fixed bug which was preventing out params of certain types
from being received; thanks a ton to huddlej at wwu.edu !
[ticket:1265]
0.4.8
=====
- orm
- Fixed bug regarding inherit_condition passed
with "A=B" versus "B=A" leading to errors
[ticket:1039]
- Changes made to new, dirty and deleted
collections in
SessionExtension.before_flush() will take
effect for that flush.
- Added label() method to InstrumentedAttribute
to establish forwards compatibility with 0.5.
- sql
- column.in_(someselect) can now be used as
a columns-clause expression without the subquery
bleeding into the FROM clause [ticket:1074]
- mysql
- Added MSMediumInteger type [ticket:1146].
- sqlite
- Supplied a custom strftime() function which
handles dates before 1900. [ticket:968]
- String's (and Unicode's, UnicodeText's, etc.)
convert_unicode logic disabled in the sqlite dialect,
to adjust for pysqlite 2.5.0's new requirement that
only Python unicode objects are accepted;
http://itsystementwicklung.de/pipermail/list-pysqlite/2008-March/000018.html
- oracle
- has_sequence() now takes schema name into account
[ticket:1155]
- added BFILE to the list of reflected types
[ticket:1121]
0.4.7p1
=====
- orm
- Added "add()" and "add_all()" to scoped_session
methods. Workaround for 0.4.7:
from sqlalchemy.orm.scoping import ScopedSession,\
instrument
setattr(
ScopedSession, "add", instrument("add"))
setattr(
ScopedSession, "add_all", instrument("add_all"))
- Fixed non-2.3 compatible usage of set() and generator
expression within relation().
0.4.7
=====
- orm
- The contains() operator when used with many-to-many
will alias() the secondary (association) table so
that multiple contains() calls will not conflict
with each other [ticket:1058]
- fixed bug preventing merge() from functioning in
conjunction with a comparable_property()
- the enable_typechecks=False setting on relation()
now only allows subtypes with inheriting mappers.
Totally unrelated types, or subtypes not set up with
mapper inheritance against the target mapper are
still not allowed.
- Added is_active flag to Sessions to detect when
a transaction is in progress [ticket:976]. This
flag is always True with a "transactional"
(in 0.5 a non-"autocommit") Session.
- sql
- Fixed bug when calling select([literal('foo')])
or select([bindparam('foo')]).
- schema
- create_all(), drop_all(), create(), drop() all raise
an error if the table name or schema name contains
more characters than that dialect's configured
character limit. Some DB's can handle too-long
table names during usage, and SQLA can handle this
as well. But various reflection/
checkfirst-during-create scenarios fail since we are
looking for the name within the DB's catalog tables.
[ticket:571]
- The index name generated when you say "index=True"
on a Column is truncated to the length appropriate
for the dialect. Additionally, an Index with a too-
long name cannot be explicitly dropped with
Index.drop(), similar to [ticket:571].
[ticket:820]
- postgres
- Repaired server_side_cursors to properly detect
text() clauses.
- Added PGCidr type. [ticket:1092]
- mysql
- Added 'CALL' to the list of SQL keywords which return
result rows.
- oracle
- Oracle get_default_schema_name() "normalizes" the name
before returning, meaning it returns a lower-case name
when the identifier is detected as case insensitive.
- creating/dropping tables takes schema name into account
when searching for the existing table, so that tables
in other owner namespaces with the same name do not
conflict [ticket:709]
- Cursors now have "arraysize" set to 50 by default on
them, the value of which is configurable using the
"arraysize" argument to create_engine() with the
Oracle dialect. This to account for cx_oracle's default
setting of "1", which has the effect of many round trips
being sent to Oracle. This actually works well in
conjunction with BLOB/CLOB-bound cursors, of which
there are any number available but only for the life of
that row request (so BufferedColumnRow is still needed,
but less so). [ticket:1062]
- sqlite
- add SLFloat type, which matches the SQLite REAL
type affinity. Previously, only SLNumeric was provided
which fulfills NUMERIC affinity, but that's not the
same as REAL.
0.4.6
=====
- orm
- Fix to the recent relation() refactoring which fixes
exotic viewonly relations which join between local and
remote table multiple times, with a common column shared
between the joins.
- Also re-established viewonly relation() configurations
that join across multiple tables.
- Added experimental relation() flag to help with
primaryjoins across functions, etc.,
_local_remote_pairs=[tuples]. This complements a complex
primaryjoin condition allowing you to provide the
individual column pairs which comprise the relation's
local and remote sides. Also improved lazy load SQL
generation to handle placing bind params inside of
functions and other expressions. (partial progress
towards [ticket:610])
- repaired single table inheritance such that you
can single-table inherit from a joined-table inherting
mapper without issue [ticket:1036].
- Fixed "concatenate tuple" bug which could occur with
Query.order_by() if clause adaption had taken place.
[ticket:1027]
- Removed ancient assertion that mapped selectables require
"alias names" - the mapper creates its own alias now if
none is present. Though in this case you need to use the
class, not the mapped selectable, as the source of column
attributes - so a warning is still issued.
- fixes to the "exists" function involving inheritance (any(),
has(), ~contains()); the full target join will be rendered
into the EXISTS clause for relations that link to subclasses.
- restored usage of append_result() extension method for primary
query rows, when the extension is present and only a single-
entity result is being returned.
- Also re-established viewonly relation() configurations that
join across multiple tables.
- removed ancient assertion that mapped selectables require
"alias names" - the mapper creates its own alias now if
none is present. Though in this case you need to use
the class, not the mapped selectable, as the source of
column attributes - so a warning is still issued.
- refined mapper._save_obj() which was unnecessarily calling
__ne__() on scalar values during flush [ticket:1015]
- added a feature to eager loading whereby subqueries set
as column_property() with explicit label names (which is not
necessary, btw) will have the label anonymized when
the instance is part of the eager join, to prevent
conflicts with a subquery or column of the same name
on the parent object. [ticket:1019]
- set-based collections |=, -=, ^= and &= are stricter about
their operands and only operate on sets, frozensets or
subclasses of the collection type. Previously, they would
accept any duck-typed set.
- added an example dynamic_dict/dynamic_dict.py, illustrating
a simple way to place dictionary behavior on top of
a dynamic_loader.
- declarative extension
- Joined table inheritance mappers use a slightly relaxed
function to create the "inherit condition" to the parent
table, so that other foreign keys to not-yet-declared
Table objects don't trigger an error.
- fixed reentrant mapper compile hang when
a declared attribute is used within ForeignKey,
ie. ForeignKey(MyOtherClass.someattribute)
- sql
- Added COLLATE support via the .collate(<collation>)
expression operator and collate(<expr>, <collation>) sql
function.
- Fixed bug with union() when applied to non-Table connected
select statements
- improved behavior of text() expressions when used as
FROM clauses, such as select().select_from(text("sometext"))
[ticket:1014]
- Column.copy() respects the value of "autoincrement",
fixes usage with Migrate [ticket:1021]
- engines
- Pool listeners can now be provided as a dictionary of
callables or a (possibly partial) duck-type of
PoolListener, your choice.
- added "rollback_returned" option to Pool which will
disable the rollback() issued when connections are
returned. This flag is only safe to use with a database
which does not support transactions (i.e. MySQL/MyISAM).
- ext
- set-based association proxies |=, -=, ^= and &= are
stricter about their operands and only operate on sets,
frozensets or other association proxies. Previously, they
would accept any duck-typed set.
- mssql
- Added "odbc_autotranslate" parameter to engine / dburi
parameters. Any given string will be passed through to the
ODBC connection string as:
"AutoTranslate=%s" % odbc_autotranslate
[ticket:1005]
- Added "odbc_options" parameter to engine / dburi
parameters. The given string is simply appended to the
SQLAlchemy-generated odbc connection string.
This should obviate the need of adding a myriad of ODBC
options in the future.
- firebird
- Handle the "SUBSTRING(:string FROM :start FOR :length)"
builtin.
0.4.5
=====
- orm
- A small change in behavior to session.merge() - existing
objects are checked for based on primary key attributes, not
necessarily _instance_key. So the widely requested
capability, that:
x = MyObject(id=1)
x = sess.merge(x)
will in fact load MyObject with id #1 from the database if
present, is now available. merge() still copies the state
of the given object to the persistent one, so an example
like the above would typically have copied "None" from all
attributes of "x" onto the persistent copy. These can be
reverted using session.expire(x).
- Also fixed behavior in merge() whereby collection elements
present on the destination but not the merged collection
were not being removed from the destination.
- Added a more aggressive check for "uncompiled mappers",
helps particularly with declarative layer [ticket:995]
- The methodology behind "primaryjoin"/"secondaryjoin" has
been refactored. Behavior should be slightly more
intelligent, primarily in terms of error messages which
have been pared down to be more readable. In a slight
number of scenarios it can better resolve the correct
foreign key than before.
- Added comparable_property(), adds query Comparator
behavior to regular, unmanaged Python properties
- the functionality of query.with_polymorphic() has
been added to mapper() as a configuration option.
It's set via several forms:
with_polymorphic='*'
with_polymorphic=[mappers]
with_polymorphic=('*', selectable)
with_polymorphic=([mappers], selectable)
This controls the default polymorphic loading strategy
for inherited mappers. When a selectable is not given,
outer joins are created for all joined-table inheriting
mappers requested. Note that the auto-create of joins
is not compatible with concrete table inheritance.
The existing select_table flag on mapper() is now
deprecated and is synonymous with
with_polymorphic('*', select_table). Note that the
underlying "guts" of select_table have been
completely removed and replaced with the newer,
more flexible approach.
The new approach also automatically allows eager loads
to work for subclasses, if they are present, for
example
sess.query(Company).options(
eagerload_all(
[Company.employees.of_type(Engineer), 'machines']
))
to load Company objects, their employees, and the
'machines' collection of employees who happen to be
Engineers. A "with_polymorphic" Query option should be
introduced soon as well which would allow per-Query
control of with_polymorphic() on relations.
- added two "experimental" features to Query,
"experimental" in that their specific name/behavior
is not carved in stone just yet: _values() and
_from_self(). We'd like feedback on these.
- _values(*columns) is given a list of column
expressions, and returns a new Query that only
returns those columns. When evaluated, the return
value is a list of tuples just like when using
add_column() or add_entity(), the only difference is
that "entity zero", i.e. the mapped class, is not
included in the results. This means it finally makes
sense to use group_by() and having() on Query, which
have been sitting around uselessly until now.
A future change to this method may include that its
ability to join, filter and allow other options not
related to a "resultset" are removed, so the feedback
we're looking for is how people want to use
_values()...i.e. at the very end, or do people prefer
to continue generating after it's called.
- _from_self() compiles the SELECT statement for the
Query (minus any eager loaders), and returns a new
Query that selects from that SELECT. So basically you
can query from a Query without needing to extract the
SELECT statement manually. This gives meaning to
operations like query[3:5]._from_self().filter(some
criterion). There's not much controversial here
except that you can quickly create highly nested
queries that are less efficient, and we want feedback
on the naming choice.
- query.order_by() and query.group_by() will accept
multiple arguments using *args (like select()
already does).
- Added some convenience descriptors to Query:
query.statement returns the full SELECT construct,
query.whereclause returns just the WHERE part of the
SELECT construct.
- Fixed/covered case when using a False/0 value as a
polymorphic discriminator.
- Fixed bug which was preventing synonym() attributes from
being used with inheritance
- Fixed SQL function truncation of trailing underscores
[ticket:996]
- When attributes are expired on a pending instance, an
error will not be raised when the "refresh" action is
triggered and no result is found.
- Session.execute can now find binds from metadata
- Adjusted the definition of "self-referential" to be any
two mappers with a common parent (this affects whether or
not aliased=True is required when joining with Query).
- Made some fixes to the "from_joinpoint" argument to
query.join() so that if the previous join was aliased and
this one isn't, the join still happens successfully.
- Assorted "cascade deletes" fixes:
- Fixed "cascade delete" operation of dynamic relations,
which had only been implemented for foreign-key
nulling behavior in 0.4.2 and not actual cascading
deletes [ticket:895]
- Delete cascade without delete-orphan cascade on a
many-to-one will not delete orphans which were
disconnected from the parent before session.delete()
is called on the parent (one-to-many already had
this).
- Delete cascade with delete-orphan will delete orphans
whether or not it remains attached to its also-deleted
parent.
- delete-orphan casacde is properly detected on relations
that are present on superclasses when using inheritance.
- Fixed order_by calculation in Query to properly alias
mapper-config'ed order_by when using select_from()
- Refactored the diffing logic that kicks in when replacing
one collection with another into collections.bulk_replace,
useful to anyone building multi-level collections.
- Cascade traversal algorithm converted from recursive to
iterative to support deep object graphs.
- sql
- schema-qualified tables now will place the schemaname
ahead of the tablename in all column expressions as well
as when generating column labels. This prevents cross-
schema name collisions in all cases [ticket:999]
- can now allow selects which correlate all FROM clauses
and have no FROM themselves. These are typically
used in a scalar context, i.e. SELECT x, (SELECT x WHERE y)
FROM table. Requires explicit correlate() call.
- 'name' is no longer a required constructor argument for
Column(). It (and .key) may now be deferred until the
column is added to a Table.
- like(), ilike(), contains(), startswith(), endswith() take
an optional keyword argument "escape=<somestring>", which
is set as the escape character using the syntax "x LIKE y
ESCAPE '<somestring>'" [ticket:993], [ticket:791]
- random() is now a generic sql function and will compile to
the database's random implementation, if any.
- update().values() and insert().values() take keyword
arguments.
- Fixed an issue in select() regarding its generation of
FROM clauses, in rare circumstances two clauses could be
produced when one was intended to cancel out the other.
Some ORM queries with lots of eager loads might have seen
this symptom.
- The case() function now also takes a dictionary as its
whens parameter. It also interprets the "THEN"
expressions as values by default, meaning case([(x==y,
"foo")]) will interpret "foo" as a bound value, not a SQL
expression. use text(expr) for literal SQL expressions in
this case. For the criterion itself, these may be literal
strings only if the "value" keyword is present, otherwise
SA will force explicit usage of either text() or
literal().
- oracle
- The "owner" keyword on Table is now deprecated, and is
exactly synonymous with the "schema" keyword. Tables can
now be reflected with alternate "owner" attributes,
explicitly stated on the Table object or not using
"schema".
- All of the "magic" searching for synonyms, DBLINKs etc.
during table reflection are disabled by default unless you
specify "oracle_resolve_synonyms=True" on the Table
object. Resolving synonyms necessarily leads to some
messy guessing which we'd rather leave off by default.
When the flag is set, tables and related tables will be
resolved against synonyms in all cases, meaning if a
synonym exists for a particular table, reflection will use
it when reflecting related tables. This is stickier
behavior than before which is why it's off by default.
- declarative extension
- The "synonym" function is now directly usable with
"declarative". Pass in the decorated property using the
"descriptor" keyword argument, e.g.: somekey =
synonym('_somekey', descriptor=property(g, s))
- The "deferred" function is usable with "declarative".
Simplest usage is to declare deferred and Column together,
e.g.: data = deferred(Column(Text))
- Declarative also gained @synonym_for(...) and
@comparable_using(...), front-ends for synonym and
comparable_property.
- Improvements to mapper compilation when using declarative;
already-compiled mappers will still trigger compiles of
other uncompiled mappers when used [ticket:995]
- Declarative will complete setup for Columns lacking names,
allows a more DRY syntax.
class Foo(Base):
__tablename__ = 'foos'
id = Column(Integer, primary_key=True)
- inheritance in declarative can be disabled when sending
"inherits=None" to __mapper_args__.
- declarative_base() takes optional kwarg "mapper", which
is any callable/class/method that produces a mapper,
such as declarative_base(mapper=scopedsession.mapper).
This property can also be set on individual declarative
classes using the "__mapper_cls__" property.
- postgres
- Got PG server side cursors back into shape, added fixed
unit tests as part of the default test suite. Added
better uniqueness to the cursor ID [ticket:1001]
- oracle
- The "owner" keyword on Table is now deprecated, and is
exactly synonymous with the "schema" keyword. Tables can
now be reflected with alternate "owner" attributes,
explicitly stated on the Table object or not using
"schema".
- All of the "magic" searching for synonyms, DBLINKs etc.
during table reflection are disabled by default unless you
specify "oracle_resolve_synonyms=True" on the Table
object. Resolving synonyms necessarily leads to some
messy guessing which we'd rather leave off by default.
When the flag is set, tables and related tables will be
resolved against synonyms in all cases, meaning if a
synonym exists for a particular table, reflection will use
it when reflecting related tables. This is stickier
behavior than before which is why it's off by default.
- mssql
- Reflected tables will now automatically load other tables
which are referenced by Foreign keys in the auto-loaded
table, [ticket:979].
- Added executemany check to skip identity fetch, [ticket:916].
- Added stubs for small date type, [ticket:884]
- Added a new 'driver' keyword parameter for the pyodbc dialect.
Will substitute into the ODBC connection string if given,
defaults to 'SQL Server'.
- Added a new 'max_identifier_length' keyword parameter for
the pyodbc dialect.
- Improvements to pyodbc + Unix. If you couldn't get that
combination to work before, please try again.
- mysql
- The connection.info keys the dialect uses to cache server
settings have changed and are now namespaced.
0.4.4
------
- sql
- Can again create aliases of selects against textual FROM
clauses, [ticket:975]
- The value of a bindparam() can be a callable, in which
case it's evaluated at statement execution time to get the
value.
- Added exception wrapping/reconnect support to result set
fetching. Reconnect works for those databases that raise
a catchable data error during results (i.e. doesn't work
on MySQL) [ticket:978]
- Implemented two-phase API for "threadlocal" engine, via
engine.begin_twophase(), engine.prepare() [ticket:936]
- Fixed bug which was preventing UNIONS from being
cloneable, [ticket:986]
- Added "bind" keyword argument to insert(), update(),
delete() and DDL(). The .bind property is now assignable
on those statements as well as on select().
- Insert statements can now be compiled with extra "prefix"
words between INSERT and INTO, for vendor extensions like
MySQL's INSERT IGNORE INTO table.
- orm
- any(), has(), contains(), ~contains(), attribute level ==
and != now work properly with self-referential relations -
the clause inside the EXISTS is aliased on the "remote"
side to distinguish it from the parent table. This
applies to single table self-referential as well as
inheritance-based self-referential.
- Repaired behavior of == and != operators at the relation()
level when compared against NULL for one-to-one relations
[ticket:985]
- Fixed bug whereby session.expire() attributes were not
loading on an polymorphically-mapped instance mapped by a
select_table mapper.
- Added query.with_polymorphic() - specifies a list of
classes which descend from the base class, which will be
added to the FROM clause of the query. Allows subclasses
to be used within filter() criterion as well as eagerly
loads the attributes of those subclasses.
- Your cries have been heard: removing a pending item from
an attribute or collection with delete-orphan expunges the
item from the session; no FlushError is raised. Note that
if you session.save()'ed the pending item explicitly, the
attribute/collection removal still knocks it out.
- session.refresh() and session.expire() raise an error when
called on instances which are not persistent within the
session
- Fixed potential generative bug when the same Query was
used to generate multiple Query objects using join().
- Fixed bug which was introduced in 0.4.3, whereby loading
an already-persistent instance mapped with joined table
inheritance would trigger a useless "secondary" load from
its joined table, when using the default "select"
polymorphic_fetch. This was due to attributes being
marked as expired during its first load and not getting
unmarked from the previous "secondary" load. Attributes
are now unexpired based on presence in __dict__ after any
load or commit operation succeeds.
- Deprecated Query methods apply_sum(), apply_max(),
apply_min(), apply_avg(). Better methodologies are
coming....
- relation() can accept a callable for its first argument,
which returns the class to be related. This is in place
to assist declarative packages to define relations without
classes yet being in place.
- Added a new "higher level" operator called "of_type()":
used in join() as well as with any() and has(), qualifies
the subclass which will be used in filter criterion, e.g.:
query.filter(Company.employees.of_type(Engineer).
any(Engineer.name=='foo'))
or
query.join(Company.employees.of_type(Engineer)).
filter(Engineer.name=='foo')
- Preventive code against a potential lost-reference bug in
flush().
- Expressions used in filter(), filter_by() and others, when
they make usage of a clause generated from a relation
using the identity of a child object (e.g.,
filter(Parent.child==<somechild>)), evaluate the actual
primary key value of <somechild> at execution time so that
the autoflush step of the Query can complete, thereby
populating the PK value of <somechild> in the case that
<somechild> was pending.
- setting the relation()-level order by to a column in the
many-to-many "secondary" table will now work with eager
loading, previously the "order by" wasn't aliased against
the secondary table's alias.
- Synonyms riding on top of existing descriptors are now
full proxies to those descriptors.
- dialects
- Invalid SQLite connection URLs now raise an error.
- postgres TIMESTAMP renders correctly [ticket:981]
- postgres PGArray is a "mutable" type by default; when used
with the ORM, mutable-style equality/ copy-on-write
techniques are used to test for changes.
- extensions
- a new super-small "declarative" extension has been added,
which allows Table and mapper() configuration to take
place inline underneath a class declaration. This
extension differs from ActiveMapper and Elixir in that it
does not redefine any SQLAlchemy semantics at all; literal
Column, Table and relation() constructs are used to define
the class behavior and table definition.
0.4.3
------
- sql
- Added "schema.DDL", an executable free-form DDL statement.
DDLs can be executed in isolation or attached to Table or
MetaData instances and executed automatically when those
objects are created and/or dropped.
- Table columns and constraints can be overridden on a an
existing table (such as a table that was already reflected)
using the 'useexisting=True' flag, which now takes into
account the arguments passed along with it.
- Added a callable-based DDL events interface, adds hooks
before and after Tables and MetaData create and drop.
- Added generative where(<criterion>) method to delete() and
update() constructs which return a new object with criterion
joined to existing criterion via AND, just like
select().where().
- Added "ilike()" operator to column operations. Compiles to
ILIKE on postgres, lower(x) LIKE lower(y) on all
others. [ticket:727]
- Added "now()" as a generic function; on SQLite, Oracle
and MSSQL compiles as "CURRENT_TIMESTAMP"; "now()" on
all others. [ticket:943]
- The startswith(), endswith(), and contains() operators now
concatenate the wildcard operator with the given operand in
SQL, i.e. "'%' || <bindparam>" in all cases, accept
text('something') operands properly [ticket:962]
- cast() accepts text('something') and other non-literal
operands properly [ticket:962]
- fixed bug in result proxy where anonymously generated
column labels would not be accessible using their straight
string name
- Deferrable constraints can now be defined.
- Added "autocommit=True" keyword argument to select() and
text(), as well as generative autocommit() method on
select(); for statements which modify the database through
some user-defined means other than the usual INSERT/UPDATE/
DELETE etc. This flag will enable "autocommit" behavior
during execution if no transaction is in progress.
[ticket:915]
- The '.c.' attribute on a selectable now gets an entry for
every column expression in its columns clause. Previously,
"unnamed" columns like functions and CASE statements weren't
getting put there. Now they will, using their full string
representation if no 'name' is available.
- a CompositeSelect, i.e. any union(), union_all(),
intersect(), etc. now asserts that each selectable contains
the same number of columns. This conforms to the
corresponding SQL requirement.
- The anonymous 'label' generated for otherwise unlabeled
functions and expressions now propagates outwards at compile
time for expressions like select([select([func.foo()])]).
- Building on the above ideas, CompositeSelects now build up
their ".c." collection based on the names present in the
first selectable only; corresponding_column() now works
fully for all embedded selectables.
- Oracle and others properly encode SQL used for defaults like
sequences, etc., even if no unicode idents are used since
identifier preparer may return a cached unicode identifier.
- Column and clause comparisons to datetime objects on the
left hand side of the expression now work (d < table.c.col).
(datetimes on the RHS have always worked, the LHS exception
is a quirk of the datetime implementation.)
- orm
- Every Session.begin() must now be accompanied by a
corresponding commit() or rollback() unless the session is
closed with Session.close(). This also includes the begin()
which is implicit to a session created with
transactional=True. The biggest change introduced here is
that when a Session created with transactional=True raises
an exception during flush(), you must call
Session.rollback() or Session.close() in order for that
Session to continue after an exception.
- Fixed merge() collection-doubling bug when merging transient
entities with backref'ed collections. [ticket:961]
- merge(dont_load=True) does not accept transient entities,
this is in continuation with the fact that
merge(dont_load=True) does not accept any "dirty" objects
either.
- Added standalone "query" class attribute generated by a
scoped_session. This provides MyClass.query without using
Session.mapper. Use via:
MyClass.query = Session.query_property()
- The proper error message is raised when trying to access
expired instance attributes with no session present
- dynamic_loader() / lazy="dynamic" now accepts and uses
the order_by parameter in the same way in which it works
with relation().
- Added expire_all() method to Session. Calls expire() for
all persistent instances. This is handy in conjunction
with...
- Instances which have been partially or fully expired will
have their expired attributes populated during a regular
Query operation which affects those objects, preventing a
needless second SQL statement for each instance.
- Dynamic relations, when referenced, create a strong
reference to the parent object so that the query still has a
parent to call against even if the parent is only created
(and otherwise dereferenced) within the scope of a single
expression. [ticket:938]
- Added a mapper() flag "eager_defaults". When set to True,
defaults that are generated during an INSERT or UPDATE
operation are post-fetched immediately, instead of being
deferred until later. This mimics the old 0.3 behavior.
- query.join() can now accept class-mapped attributes as
arguments. These can be used in place or in any combination
with strings. In particular this allows construction of
joins to subclasses on a polymorphic relation, i.e.:
query(Company).join(['employees', Engineer.name])
- query.join() can also accept tuples of attribute name/some
selectable as arguments. This allows construction of joins
*from* subclasses of a polymorphic relation, i.e.:
query(Company).\
join(
[('employees', people.join(engineer)), Engineer.name]
)
- General improvements to the behavior of join() in
conjunction with polymorphic mappers, i.e. joining from/to
polymorphic mappers and properly applying aliases.
- Fixed/improved behavior when a mapper determines the natural
"primary key" of a mapped join, it will more effectively
reduce columns which are equivalent via foreign key
relation. This affects how many arguments need to be sent
to query.get(), among other things. [ticket:933]
- The lazy loader can now handle a join condition where the
"bound" column (i.e. the one that gets the parent id sent as
a bind parameter) appears more than once in the join
condition. Specifically this allows the common task of a
relation() which contains a parent-correlated subquery, such
as "select only the most recent child item". [ticket:946]
- Fixed bug in polymorphic inheritance where an incorrect
exception is raised when base polymorphic_on column does not
correspond to any columns within the local selectable of an
inheriting mapper more than one level deep
- Fixed bug in polymorphic inheritance which made it difficult
to set a working "order_by" on a polymorphic mapper.
- Fixed a rather expensive call in Query that was slowing down
polymorphic queries.
- "Passive defaults" and other "inline" defaults can now be
loaded during a flush() call if needed; in particular, this
allows constructing relations() where a foreign key column
references a server-side-generated, non-primary-key
column. [ticket:954]
- Additional Session transaction fixes/changes:
- Fixed bug with session transaction management: parent
transactions weren't started on the connection when
adding a connection to a nested transaction.
- session.transaction now always refers to the innermost
active transaction, even when commit/rollback are called
directly on the session transaction object.
- Two-phase transactions can now be prepared.
- When preparing a two-phase transaction fails on one
connection, all the connections are rolled back.
- session.close() didn't close all transactions when
nested transactions were used.
- rollback() previously erroneously set the current
transaction directly to the parent of the transaction
that could be rolled back to. Now it rolls back the next
transaction up that can handle it, but sets the current
transaction to it's parent and inactivates the
transactions in between. Inactive transactions can only
be rolled back or closed, any other call results in an
error.
- autoflush for commit() wasn't flushing for simple
subtransactions.
- unitofwork flush didn't close the failed transaction
when the session was not in a transaction and commiting
the transaction failed.
- Miscellaneous tickets: [ticket:940] [ticket:964]
- general
- Fixed a variety of hidden and some not-so-hidden
compatibility issues for Python 2.3, thanks to new support
for running the full test suite on 2.3.
- Warnings are now issued as type exceptions.SAWarning.
- dialects
- Better support for schemas in SQLite (linked in by ATTACH
DATABASE ... AS name). In some cases in the past, schema
names were ommitted from generated SQL for SQLite. This is
no longer the case.
- table_names on SQLite now picks up temporary tables as well.
- Auto-detect an unspecified MySQL ANSI_QUOTES mode during
reflection operations, support for changing the mode
midstream. Manual mode setting is still required if no
reflection is used.
- Fixed reflection of TIME columns on SQLite.
- Finally added PGMacAddr type to postgres [ticket:580]
- Reflect the sequence associated to a PK field (typically
with a BEFORE INSERT trigger) under Firebird
- Oracle assembles the correct columns in the result set
column mapping when generating a LIMIT/OFFSET subquery,
allows columns to map properly to result sets even if
long-name truncation kicks in [ticket:941]
- MSSQL now includes EXEC in the _is_select regexp, which
should allow row-returning stored procedures to be used.
- MSSQL now includes an experimental implementation of
LIMIT/OFFSET using the ANSI SQL row_number() function, so it
requires MSSQL-2005 or higher. To enable the feature, add
"has_window_funcs" to the keyword arguments for connect, or
add "?has_window_funcs=1" to your dburi query arguments.
- ext
- Changed ext.activemapper to use a non-transactional session
for the objectstore.
- Fixed output order of "['a'] + obj.proxied" binary operation
on association-proxied lists.
0.4.2p3
------
- general
- sub version numbering scheme changed to suite
setuptools version number rules; easy_install -u
should now get this version over 0.4.2.
- sql
- Text type is properly exported now and does not
raise a warning on DDL create; String types with no
length only raise warnings during CREATE TABLE
[ticket:912]
- new UnicodeText type is added, to specify an
encoded, unlengthed Text type
- fixed bug in union() so that select() statements
which don't derive from FromClause objects can be
unioned
- orm
- fixed bug with session.dirty when using "mutable
scalars" (such as PickleTypes)
- added a more descriptive error message when flushing
on a relation() that has non-locally-mapped columns
in its primary or secondary join condition
- dialects
- Fixed reflection of mysql empty string column
defaults.
0.4.2b (0.4.2p2)
------
- sql
- changed name of TEXT to Text since its a "generic"
type; TEXT name is deprecated until 0.5. The
"upgrading" behavior of String to Text when no
length is present is also deprecated until 0.5; will
issue a warning when used for CREATE TABLE
statements (String with no length for SQL expression
purposes is still fine) [ticket:912]
- generative select.order_by(None) / group_by(None)
was not managing to reset order by/group by
criterion, fixed [ticket:924]
- orm
- suppressing *all* errors in
InstanceState.__cleanup() now.
- fixed an attribute history bug whereby assigning a
new collection to a collection-based attribute which
already had pending changes would generate incorrect
history [ticket:922]
- fixed delete-orphan cascade bug whereby setting the
same object twice to a scalar attribute could log it
as an orphan [ticket:925]
- Fixed cascades on a += assignment to a list-based
relation.
- synonyms can now be created against props that don't
exist yet, which are later added via add_property().
This commonly includes backrefs. (i.e. you can make
synonyms for backrefs without worrying about the
order of operations) [ticket:919]
- fixed bug which could occur with polymorphic "union"
mapper which falls back to "deferred" loading of
inheriting tables
- the "columns" collection on a mapper/mapped class
(i.e. 'c') is against the mapped table, not the
select_table in the case of polymorphic "union"
loading (this shouldn't be noticeable).
- ext
- '+', '*', '+=' and '*=' support for association
proxied lists.
- dialects
- mssql - narrowed down the test for "date"/"datetime"
in MSDate/ MSDateTime subclasses so that incoming
"datetime" objects don't get mis-interpreted as
"date" objects and vice versa, [ticket:923]
0.4.2a (0.4.2p1)
------
- orm
- fixed fairly critical bug whereby the same instance could be listed
more than once in the unitofwork.new collection; most typically
reproduced when using a combination of inheriting mappers and
ScopedSession.mapper, as the multiple __init__ calls per instance
could save() the object with distinct _state objects
- added very rudimentary yielding iterator behavior to Query. Call
query.yield_per(<number of rows>) and evaluate the Query in an
iterative context; every collection of N rows will be packaged up
and yielded. Use this method with extreme caution since it does
not attempt to reconcile eagerly loaded collections across
result batch boundaries, nor will it behave nicely if the same
instance occurs in more than one batch. This means that an eagerly
loaded collection will get cleared out if it's referenced in more than
one batch, and in all cases attributes will be overwritten on instances
that occur in more than one batch.
- Fixed in-place set mutation operators for set collections and association
proxied sets. [ticket:920]
- dialects
- Fixed the missing call to subtype result processor for the PGArray
type. [ticket:913]
0.4.2
-----
- sql
- generic functions ! we introduce a database of known SQL functions, such
as current_timestamp, coalesce, and create explicit function objects
representing them. These objects have constrained argument lists, are
type aware, and can compile in a dialect-specific fashion. So saying
func.char_length("foo", "bar") raises an error (too many args),
func.coalesce(datetime.date(2007, 10, 5), datetime.date(2005, 10, 15))
knows that its return type is a Date. We only have a few functions
represented so far but will continue to add to the system [ticket:615]
- auto-reconnect support improved; a Connection can now automatically
reconnect after its underlying connection is invalidated, without
needing to connect() again from the engine. This allows an ORM session
bound to a single Connection to not need a reconnect.
Open transactions on the Connection must be rolled back after an invalidation
of the underlying connection else an error is raised. Also fixed
bug where disconnect detect was not being called for cursor(), rollback(),
or commit().
- added new flag to String and create_engine(),
assert_unicode=(True|False|'warn'|None). Defaults to `False` or `None` on
create_engine() and String, `'warn'` on the Unicode type. When `True`,
results in all unicode conversion operations raising an exception when a
non-unicode bytestring is passed as a bind parameter. 'warn' results
in a warning. It is strongly advised that all unicode-aware applications
make proper use of Python unicode objects (i.e. u'hello' and not 'hello')
so that data round trips accurately.
- generation of "unique" bind parameters has been simplified to use the same
"unique identifier" mechanisms as everything else. This doesn't affect
user code, except any code that might have been hardcoded against the generated
names. Generated bind params now have the form "<paramname>_<num>",
whereas before only the second bind of the same name would have this form.
- select().as_scalar() will raise an exception if the select does not have
exactly one expression in its columns clause.
- bindparam() objects themselves can be used as keys for execute(), i.e.
statement.execute({bind1:'foo', bind2:'bar'})
- added new methods to TypeDecorator, process_bind_param() and
process_result_value(), which automatically take advantage of the processing
of the underlying type. Ideal for using with Unicode or Pickletype.
TypeDecorator should now be the primary way to augment the behavior of any
existing type including other TypeDecorator subclasses such as PickleType.
- selectables (and others) will issue a warning when two columns in
their exported columns collection conflict based on name.
- tables with schemas can still be used in sqlite, firebird,
schema name just gets dropped [ticket:890]
- changed the various "literal" generation functions to use an anonymous
bind parameter. not much changes here except their labels now look
like ":param_1", ":param_2" instead of ":literal"
- column labels in the form "tablename.columname", i.e. with a dot, are now
supported.
- from_obj keyword argument to select() can be a scalar or a list.
- orm
- a major behavioral change to collection-based backrefs: they no
longer trigger lazy loads ! "reverse" adds and removes
are queued up and are merged with the collection when it is
actually read from and loaded; but do not trigger a load beforehand.
For users who have noticed this behavior, this should be much more
convenient than using dynamic relations in some cases; for those who
have not, you might notice your apps using a lot fewer queries than
before in some situations. [ticket:871]
- mutable primary key support is added. primary key columns can be
changed freely, and the identity of the instance will change upon
flush. In addition, update cascades of foreign key referents (primary
key or not) along relations are supported, either in tandem with the
database's ON UPDATE CASCADE (required for DB's like Postgres) or
issued directly by the ORM in the form of UPDATE statements, by setting
the flag "passive_cascades=False".
- inheriting mappers now inherit the MapperExtensions of their parent
mapper directly, so that all methods for a particular MapperExtension
are called for subclasses as well. As always, any MapperExtension
can return either EXT_CONTINUE to continue extension processing
or EXT_STOP to stop processing. The order of mapper resolution is:
<extensions declared on the classes mapper> <extensions declared on the
classes' parent mapper> <globally declared extensions>.
Note that if you instantiate the same extension class separately
and then apply it individually for two mappers in the same inheritance
chain, the extension will be applied twice to the inheriting class,
and each method will be called twice.
To apply a mapper extension explicitly to each inheriting class but
have each method called only once per operation, use the same
instance of the extension for both mappers.
[ticket:490]
- MapperExtension.before_update() and after_update() are now called
symmetrically; previously, an instance that had no modified column
attributes (but had a relation() modification) could be called with
before_update() but not after_update() [ticket:907]
- columns which are missing from a Query's select statement
now get automatically deferred during load.
- mapped classes which extend "object" and do not provide an
__init__() method will now raise TypeError if non-empty *args
or **kwargs are present at instance construction time (and are
not consumed by any extensions such as the scoped_session mapper),
consistent with the behavior of normal Python classes [ticket:908]
- fixed Query bug when filter_by() compares a relation against None
[ticket:899]
- improved support for pickling of mapped entities. Per-instance
lazy/deferred/expired callables are now serializable so that
they serialize and deserialize with _state.
- new synonym() behavior: an attribute will be placed on the mapped
class, if one does not exist already, in all cases. if a property
already exists on the class, the synonym will decorate the property
with the appropriate comparison operators so that it can be used in in
column expressions just like any other mapped attribute (i.e. usable in
filter(), etc.) the "proxy=True" flag is deprecated and no longer means
anything. Additionally, the flag "map_column=True" will automatically
generate a ColumnProperty corresponding to the name of the synonym,
i.e.: 'somename':synonym('_somename', map_column=True) will map the
column named 'somename' to the attribute '_somename'. See the example
in the mapper docs. [ticket:801]
- Query.select_from() now replaces all existing FROM criterion with
the given argument; the previous behavior of constructing a list
of FROM clauses was generally not useful as is required
filter() calls to create join criterion, and new tables introduced
within filter() already add themselves to the FROM clause. The
new behavior allows not just joins from the main table, but select
statements as well. Filter criterion, order bys, eager load
clauses will be "aliased" against the given statement.
- this month's refactoring of attribute instrumentation changes
the "copy-on-load" behavior we've had since midway through 0.3
with "copy-on-modify" in most cases. This takes a sizable chunk
of latency out of load operations and overall does less work
as only attributes which are actually modified get their
"committed state" copied. Only "mutable scalar" attributes
(i.e. a pickled object or other mutable item), the reason for
the copy-on-load change in the first place, retain the old
behavior.
- a slight behavioral change to attributes is, del'ing an attribute
does *not* cause the lazyloader of that attribute to fire off again;
the "del" makes the effective value of the attribute "None". To
re-trigger the "loader" for an attribute, use
session.expire(instance, [attrname]).
- query.filter(SomeClass.somechild == None), when comparing
a many-to-one property to None, properly generates "id IS NULL"
including that the NULL is on the right side.
- query.order_by() takes into account aliased joins, i.e.
query.join('orders', aliased=True).order_by(Order.id)
- eagerload(), lazyload(), eagerload_all() take an optional
second class-or-mapper argument, which will select the mapper
to apply the option towards. This can select among other
mappers which were added using add_entity().
- eagerloading will work with mappers added via add_entity().
- added "cascade delete" behavior to "dynamic" relations just like
that of regular relations. if passive_deletes flag (also just added)
is not set, a delete of the parent item will trigger a full load of
the child items so that they can be deleted or updated accordingly.
- also with dynamic, implemented correct count() behavior as well
as other helper methods.
- fix to cascades on polymorphic relations, such that cascades
from an object to a polymorphic collection continue cascading
along the set of attributes specific to each element in the collection.
- query.get() and query.load() do not take existing filter or other
criterion into account; these methods *always* look up the given id
in the database or return the current instance from the identity map,
disregarding any existing filter, join, group_by or other criterion
which has been configured. [ticket:893]
- added support for version_id_col in conjunction with inheriting mappers.
version_id_col is typically set on the base mapper in an inheritance
relationship where it takes effect for all inheriting mappers.
[ticket:883]
- relaxed rules on column_property() expressions having labels; any
ColumnElement is accepted now, as the compiler auto-labels non-labeled
ColumnElements now. a selectable, like a select() statement, still
requires conversion to ColumnElement via as_scalar() or label().
- fixed backref bug where you could not del instance.attr if attr
was None
- several ORM attributes have been removed or made private:
mapper.get_attr_by_column(), mapper.set_attr_by_column(),
mapper.pks_by_table, mapper.cascade_callable(),
MapperProperty.cascade_callable(), mapper.canload(),
mapper.save_obj(), mapper.delete_obj(), mapper._mapper_registry,
attributes.AttributeManager
- Assigning an incompatible collection type to a relation attribute now
raises TypeError instead of sqlalchemy's ArgumentError.
- Bulk assignment of a MappedCollection now raises an error if a key in the
incoming dictionary does not match the key that the collection's keyfunc
would use for that value. [ticket:886]
- Custom collections can now specify a @converter method to translate
objects used in "bulk" assignment into a stream of values, as in::
obj.col = [newval1, newval2]
# or
obj.dictcol = {'foo': newval1, 'bar': newval2}
The MappedCollection uses this hook to ensure that incoming key/value
pairs are sane from the collection's perspective.
- fixed endless loop issue when using lazy="dynamic" on both
sides of a bi-directional relationship [ticket:872]
- more fixes to the LIMIT/OFFSET aliasing applied with Query + eagerloads,
in this case when mapped against a select statement [ticket:904]
- fix to self-referential eager loading such that if the same mapped
instance appears in two or more distinct sets of columns in the same
result set, its eagerly loaded collection will be populated regardless
of whether or not all of the rows contain a set of "eager" columns for
that collection. this would also show up as a KeyError when fetching
results with join_depth turned on.
- fixed bug where Query would not apply a subquery to the SQL when LIMIT
was used in conjunction with an inheriting mapper where the eager
loader was only in the parent mapper.
- clarified the error message which occurs when you try to update()
an instance with the same identity key as an instance already present
in the session.
- some clarifications and fixes to merge(instance, dont_load=True).
fixed bug where lazy loaders were getting disabled on returned instances.
Also, we currently do not support merging an instance which has uncommitted
changes on it, in the case that dont_load=True is used....this will
now raise an error. This is due to complexities in merging the
"committed state" of the given instance to correctly correspond to the
newly copied instance, as well as other modified state.
Since the use case for dont_load=True is caching, the given instances
shouldn't have any uncommitted changes on them anyway.
We also copy the instances over without using any events now, so that
the 'dirty' list on the new session remains unaffected.
- fixed bug which could arise when using session.begin_nested() in conjunction
with more than one level deep of enclosing session.begin() statements
- fixed session.refresh() with instance that has custom entity_name
[ticket:914]
- dialects
- sqlite SLDate type will not erroneously render "microseconds" portion
of a datetime or time object.
- oracle
- added disconnect detection support for Oracle
- some cleanup to binary/raw types so that cx_oracle.LOB is detected
on an ad-hoc basis [ticket:902]
- MSSQL
- PyODBC no longer has a global "set nocount on".
- Fix non-identity integer PKs on autload [ticket:824]
- Better support for convert_unicode [ticket:839]
- Less strict date conversion for pyodbc/adodbapi [ticket:842]
- Schema-qualified tables / autoload [ticket:901]
- Firebird backend
- does properly reflect domains (partially fixing [ticket:410]) and
PassiveDefaults
- reverted to use default poolclass (was set to SingletonThreadPool in
0.4.0 [3562] for test purposes)
- map func.length() to 'char_length' (easily overridable with the UDF
'strlen' on old versions of Firebird)
0.4.1
-----
- sql
- the "shortname" keyword parameter on bindparam() has been
deprecated.
- Added contains operator (generates a "LIKE %<other>%" clause).
- anonymous column expressions are automatically labeled.
e.g. select([x* 5]) produces "SELECT x * 5 AS anon_1".
This allows the labelname to be present in the cursor.description
which can then be appropriately matched to result-column processing
rules. (we can't reliably use positional tracking for result-column
matches since text() expressions may represent multiple columns).
- operator overloading is now controlled by TypeEngine objects - the
one built-in operator overload so far is String types overloading
'+' to be the string concatenation operator.
User-defined types can also define their own operator overloading
by overriding the adapt_operator(self, op) method.
- untyped bind parameters on the right side of a binary expression
will be assigned the type of the left side of the operation, to better
enable the appropriate bind parameter processing to take effect
[ticket:819]
- Removed regular expression step from most statement compilations.
Also fixes [ticket:833]
- Fixed empty (zero column) sqlite inserts, allowing inserts on
autoincrementing single column tables.
- Fixed expression translation of text() clauses; this repairs various
ORM scenarios where literal text is used for SQL expressions
- Removed ClauseParameters object; compiled.params returns a regular
dictionary now, as well as result.last_inserted_params() /
last_updated_params().
- Fixed INSERT statements w.r.t. primary key columns that have
SQL-expression based default generators on them; SQL expression
executes inline as normal but will not trigger a "postfetch" condition
for the column, for those DB's who provide it via cursor.lastrowid
- func. objects can be pickled/unpickled [ticket:844]
- rewrote and simplified the system used to "target" columns across
selectable expressions. On the SQL side this is represented by the
"corresponding_column()" method. This method is used heavily by the ORM
to "adapt" elements of an expression to similar, aliased expressions,
as well as to target result set columns originally bound to a
table or selectable to an aliased, "corresponding" expression. The new
rewrite features completely consistent and accurate behavior.
- Added a field ("info") for storing arbitrary data on schema items
[ticket:573]
- The "properties" collection on Connections has been renamed "info" to
match schema's writable collections. Access is still available via
the "properties" name until 0.5.
- fixed the close() method on Transaction when using strategy='threadlocal'
- fix to compiled bind parameters to not mistakenly populate None
[ticket:853]
- <Engine|Connection>._execute_clauseelement becomes a public method
Connectable.execute_clauseelement
- orm
- eager loading with LIMIT/OFFSET applied no longer adds the primary
table joined to a limited subquery of itself; the eager loads now
join directly to the subquery which also provides the primary table's
columns to the result set. This eliminates a JOIN from all eager loads
with LIMIT/OFFSET. [ticket:843]
- session.refresh() and session.expire() now support an additional argument
"attribute_names", a list of individual attribute keynames to be refreshed
or expired, allowing partial reloads of attributes on an already-loaded
instance. [ticket:802]
- added op() operator to instrumented attributes; i.e.
User.name.op('ilike')('%somename%') [ticket:767]
- Mapped classes may now define __eq__, __hash__, and __nonzero__ methods
with arbitrary semantics. The orm now handles all mapped instances on
an identity-only basis. (e.g. 'is' vs '==') [ticket:676]
- the "properties" accessor on Mapper is removed; it now throws an informative
exception explaining the usage of mapper.get_property() and
mapper.iterate_properties
- added having() method to Query, applies HAVING to the generated statement
in the same way as filter() appends to the WHERE clause.
- The behavior of query.options() is now fully based on paths, i.e. an
option such as eagerload_all('x.y.z.y.x') will apply eagerloading to
only those paths, i.e. and not 'x.y.x'; eagerload('children.children')
applies only to exactly two-levels deep, etc. [ticket:777]
- PickleType will compare using `==` when set up with mutable=False,
and not the `is` operator. To use `is` or any other comparator, send
in a custom comparison function using PickleType(comparator=my_custom_comparator).
- query doesn't throw an error if you use distinct() and an order_by()
containing UnaryExpressions (or other) together [ticket:848]
- order_by() expressions from joined tables are properly added to columns
clause when using distinct() [ticket:786]
- fixed error where Query.add_column() would not accept a class-bound
attribute as an argument; Query also raises an error if an invalid
argument was sent to add_column() (at instances() time) [ticket:858]
- added a little more checking for garbage-collection dereferences in
InstanceState.__cleanup() to reduce "gc ignored" errors on app
shutdown
- The session API has been solidified:
- It's an error to session.save() an object which is already
persistent [ticket:840]
- It's an error to session.delete() an object which is *not*
persistent.
- session.update() and session.delete() raise an error when updating
or deleting an instance that is already in the session with a
different identity.
- The session checks more carefully when determining "object X already
in another session"; e.g. if you pickle a series of objects and
unpickle (i.e. as in a Pylons HTTP session or similar), they can go
into a new session without any conflict
- merge() includes a keyword argument "dont_load=True". setting this
flag will cause the merge operation to not load any data from the
database in response to incoming detached objects, and will accept
the incoming detached object as though it were already present in
that session. Use this to merge detached objects from external
caching systems into the session.
- Deferred column attributes no longer trigger a load operation when the
attribute is assigned to. In those cases, the newly assigned value
will be present in the flushes' UPDATE statement unconditionally.
- Fixed a truncation error when re-assigning a subset of a collection
(obj.relation = obj.relation[1:]) [ticket:834]
- De-cruftified backref configuration code, backrefs which step on
existing properties now raise an error [ticket:832]
- Improved behavior of add_property() etc., fixed [ticket:831] involving
synonym/deferred.
- Fixed clear_mappers() behavior to better clean up after itself.
- Fix to "row switch" behavior, i.e. when an INSERT/DELETE is combined
into a single UPDATE; many-to-many relations on the parent object
update properly. [ticket:841]
- Fixed __hash__ for association proxy- these collections are unhashable,
just like their mutable Python counterparts.
- Added proxying of save_or_update, __contains__ and __iter__ methods for
scoped sessions.
- fixed very hard-to-reproduce issue where by the FROM clause of Query
could get polluted by certain generative calls [ticket:852]
- dialects
- Added experimental support for MaxDB (versions >= 7.6.03.007 only).
- oracle will now reflect "DATE" as an OracleDateTime column, not
OracleDate
- added awareness of schema name in oracle table_names() function,
fixes metadata.reflect(schema='someschema') [ticket:847]
- MSSQL anonymous labels for selection of functions made deterministic
- sqlite will reflect "DECIMAL" as a numeric column.
- Made access dao detection more reliable [ticket:828]
- Renamed the Dialect attribute 'preexecute_sequences' to
'preexecute_pk_sequences'. An attribute porxy is in place for
out-of-tree dialects using the old name.
- Added test coverage for unknown type reflection. Fixed sqlite/mysql
handling of type reflection for unknown types.
- Added REAL for mysql dialect (for folks exploiting the
REAL_AS_FLOAT sql mode).
- mysql Float, MSFloat and MSDouble constructed without arguments
now produce no-argument DDL, e.g.'FLOAT'.
- misc
- Removed unused util.hash().
0.4.0
-----
- (see 0.4.0beta1 for the start of major changes against 0.3,
as well as http://www.sqlalchemy.org/trac/wiki/WhatsNewIn04 )
- Added initial Sybase support (mxODBC so far) [ticket:785]
- Added partial index support for PostgreSQL. Use the postgres_where keyword
on the Index.
- string-based query param parsing/config file parser understands
wider range of string values for booleans [ticket:817]
- backref remove object operation doesn't fail if the other-side
collection doesn't contain the item, supports noload collections
[ticket:813]
- removed __len__ from "dynamic" collection as it would require issuing
a SQL "count()" operation, thus forcing all list evaluations to issue
redundant SQL [ticket:818]
- inline optimizations added to locate_dirty() which can greatly speed up
repeated calls to flush(), as occurs with autoflush=True [ticket:816]
- The IdentifierPreprarer's _requires_quotes test is now regex based. Any
out-of-tree dialects that provide custom sets of legal_characters or
illegal_initial_characters will need to move to regexes or override
_requires_quotes.
- Firebird has supports_sane_rowcount and supports_sane_multi_rowcount set
to False due to ticket #370 (right way).
- Improvements and fixes on Firebird reflection:
. FBDialect now mimics OracleDialect, regarding case-sensitivity of TABLE and
COLUMN names (see 'case_sensitive remotion' topic on this current file).
. FBDialect.table_names() doesn't bring system tables (ticket:796).
. FB now reflects Column's nullable property correctly.
- Fixed SQL compiler's awareness of top-level column labels as used
in result-set processing; nested selects which contain the same column
names don't affect the result or conflict with result-column metadata.
- query.get() and related functions (like many-to-one lazyloading)
use compile-time-aliased bind parameter names, to prevent
name conflicts with bind parameters that already exist in the
mapped selectable.
- Fixed three- and multi-level select and deferred inheritance loading
(i.e. abc inheritance with no select_table), [ticket:795]
- Ident passed to id_chooser in shard.py always a list.
- The no-arg ResultProxy._row_processor() is now the class attribute
`_process_row`.
- Added support for returning values from inserts and updates for
PostgreSQL 8.2+. [ticket:797]
- PG reflection, upon seeing the default schema name being used explicitly
as the "schema" argument in a Table, will assume that this is the the
user's desired convention, and will explicitly set the "schema" argument
in foreign-key-related reflected tables, thus making them match only
with Table constructors that also use the explicit "schema" argument
(even though its the default schema).
In other words, SA assumes the user is being consistent in this usage.
- fixed sqlite reflection of BOOL/BOOLEAN [ticket:808]
- Added support for UPDATE with LIMIT on mysql.
- null foreign key on a m2o doesn't trigger a lazyload [ticket:803]
- oracle does not implicitly convert to unicode for non-typed result
sets (i.e. when no TypeEngine/String/Unicode type is even being used;
previously it was detecting DBAPI types and converting regardless).
should fix [ticket:800]
- fix to anonymous label generation of long table/column names [ticket:806]
- Firebird dialect now uses SingletonThreadPool as poolclass.
- Firebird now uses dialect.preparer to format sequences names
- Fixed breakage with postgres and multiple two-phase transactions. Two-phase
commits and and rollbacks didn't automatically end up with a new transaction
as the usual dbapi commits/rollbacks do. [ticket:810]
- Added an option to the _ScopedExt mapper extension to not automatically
save new objects to session on object initialization.
- fixed Oracle non-ansi join syntax
- PickleType and Interval types (on db not supporting it natively) are now
slightly faster.
- Added Float and Time types to Firebird (FBFloat and FBTime). Fixed
BLOB SUB_TYPE for TEXT and Binary types.
- Changed the API for the in_ operator. in_() now accepts a single argument
that is a sequence of values or a selectable. The old API of passing in
values as varargs still works but is deprecated.
0.4.0beta6
----------
- The Session identity map is now *weak referencing* by default, use
weak_identity_map=False to use a regular dict. The weak dict we are using
is customized to detect instances which are "dirty" and maintain a
temporary strong reference to those instances until changes are flushed.
- Mapper compilation has been reorganized such that most compilation occurs
upon mapper construction. This allows us to have fewer calls to
mapper.compile() and also to allow class-based properties to force a
compilation (i.e. User.addresses == 7 will compile all mappers; this is
[ticket:758]). The only caveat here is that an inheriting mapper now
looks for its inherited mapper upon construction; so mappers within
inheritance relationships need to be constructed in inheritance order
(which should be the normal case anyway).
- added "FETCH" to the keywords detected by Postgres to indicate a
result-row holding statement (i.e. in addition to "SELECT").
- Added full list of SQLite reserved keywords so that they get escaped
properly.
- Tightened up the relationship between the Query's generation of "eager
load" aliases, and Query.instances() which actually grabs the eagerly
loaded rows. If the aliases were not specifically generated for that
statement by EagerLoader, the EagerLoader will not take effect when the
rows are fetched. This prevents columns from being grabbed accidentally
as being part of an eager load when they were not meant for such, which
can happen with textual SQL as well as some inheritance situations. It's
particularly important since the "anonymous aliasing" of columns uses
simple integer counts now to generate labels.
- Removed "parameters" argument from clauseelement.compile(), replaced with
"column_keys". The parameters sent to execute() only interact with the
insert/update statement compilation process in terms of the column names
present but not the values for those columns. Produces more consistent
execute/executemany behavior, simplifies things a bit internally.
- Added 'comparator' keyword argument to PickleType. By default, "mutable"
PickleType does a "deep compare" of objects using their dumps()
representation. But this doesn't work for dictionaries. Pickled objects
which provide an adequate __eq__() implementation can be set up with
"PickleType(comparator=operator.eq)" [ticket:560]
- Added session.is_modified(obj) method; performs the same "history"
comparison operation as occurs within a flush operation; setting
include_collections=False gives the same result as is used when the flush
determines whether or not to issue an UPDATE for the instance's row.
- Added "schema" argument to Sequence; use this with Postgres /Oracle when
the sequence is located in an alternate schema. Implements part of
[ticket:584], should fix [ticket:761].
- Fixed reflection of the empty string for mysql enums.
- Changed MySQL dialect to use the older LIMIT <offset>, <limit> syntax
instead of LIMIT <l> OFFSET <o> for folks using 3.23. [ticket:794]
- Added 'passive_deletes="all"' flag to relation(), disables all nulling-out
of foreign key attributes during a flush where the parent object is
deleted.
- Column defaults and onupdates, executing inline, will add parenthesis for
subqueries and other parenthesis-requiring expressions
- The behavior of String/Unicode types regarding that they auto-convert to
TEXT/CLOB when no length is present now occurs *only* for an exact type of
String or Unicode with no arguments. If you use VARCHAR or NCHAR
(subclasses of String/Unicode) with no length, they will be interpreted by
the dialect as VARCHAR/NCHAR; no "magic" conversion happens there. This
is less surprising behavior and in particular this helps Oracle keep
string-based bind parameters as VARCHARs and not CLOBs [ticket:793].
- Fixes to ShardedSession to work with deferred columns [ticket:771].
- User-defined shard_chooser() function must accept "clause=None" argument;
this is the ClauseElement passed to session.execute(statement) and can be
used to determine correct shard id (since execute() doesn't take an
instance.)
- Adjusted operator precedence of NOT to match '==' and others, so that
~(x <operator> y) produces NOT (x <op> y), which is better compatible
with older MySQL versions. [ticket:764]. This doesn't apply to "~(x==y)"
as it does in 0.3 since ~(x==y) compiles to "x != y", but still applies
to operators like BETWEEN.
- Other tickets: [ticket:768], [ticket:728], [ticket:779], [ticket:757]
0.4.0beta5
----------
- Connection pool fixes; the better performance of beta4 remains but fixes
"connection overflow" and other bugs which were present (like
[ticket:754]).
- Fixed bugs in determining proper sync clauses from custom inherit
conditions. [ticket:769]
- Extended 'engine_from_config' coercion for QueuePool size / overflow.
[ticket:763]
- mysql views can be reflected again. [ticket:748]
- AssociationProxy can now take custom getters and setters.
- Fixed malfunctioning BETWEEN in orm queries.
- Fixed OrderedProperties pickling [ticket:762]
- SQL-expression defaults and sequences now execute "inline" for all
non-primary key columns during an INSERT or UPDATE, and for all columns
during an executemany()-style call. inline=True flag on any insert/update
statement also forces the same behavior with a single execute().
result.postfetch_cols() is a collection of columns for which the previous
single insert or update statement contained a SQL-side default expression.
- Fixed PG executemany() behavior, [ticket:759]
- postgres reflects tables with autoincrement=False for primary key columns
which have no defaults.
- postgres no longer wraps executemany() with individual execute() calls,
instead favoring performance. "rowcount"/"concurrency" checks with
deleted items (which use executemany) are disabled with PG since psycopg2
does not report proper rowcount for executemany().
- Tickets fixed:
- [ticket:742]
- [ticket:748]
- [ticket:760]
- [ticket:762]
- [ticket:763]
0.4.0beta4
----------
- Tidied up what ends up in your namespace when you 'from sqlalchemy import *':
- 'table' and 'column' are no longer imported. They remain available by
direct reference (as in 'sql.table' and 'sql.column') or a glob import
from the sql package. It was too easy to accidentally use a
sql.expressions.table instead of schema.Table when just starting out
with SQLAlchemy, likewise column.
- Internal-ish classes like ClauseElement, FromClause, NullTypeEngine,
etc., are also no longer imported into your namespace
- The 'Smallinteger' compatiblity name (small i!) is no longer imported,
but remains in schema.py for now. SmallInteger (big I!) is still
imported.
- The connection pool uses a "threadlocal" strategy internally to return
the same connection already bound to a thread, for "contextual" connections;
these are the connections used when you do a "connectionless" execution
like insert().execute(). This is like a "partial" version of the
"threadlocal" engine strategy but without the thread-local transaction part
of it. We're hoping it reduces connection pool overhead as well as
database usage. However, if it proves to impact stability in a negative way,
we'll roll it right back.
- Fix to bind param processing such that "False" values (like blank strings)
still get processed/encoded.
- Fix to select() "generative" behavior, such that calling column(),
select_from(), correlate(), and with_prefix() does not modify the
original select object [ticket:752]
- Added a "legacy" adapter to types, such that user-defined TypeEngine
and TypeDecorator classes which define convert_bind_param() and/or
convert_result_value() will continue to function. Also supports
calling the super() version of those methods.
- Added session.prune(), trims away instances cached in a session that
are no longer referenced elsewhere. (A utility for strong-ref
identity maps).
- Added close() method to Transaction. Closes out a transaction using
rollback if it's the outermost transaction, otherwise just ends
without affecting the outer transaction.
- Transactional and non-transactional Session integrates better with
bound connection; a close() will ensure that connection
transactional state is the same as that which existed on it before
being bound to the Session.
- Modified SQL operator functions to be module-level operators,
allowing SQL expressions to be pickleable. [ticket:735]
- Small adjustment to mapper class.__init__ to allow for Py2.6
object.__init__() behavior.
- Fixed 'prefix' argument for select()
- Connection.begin() no longer accepts nested=True, this logic is now
all in begin_nested().
- Fixes to new "dynamic" relation loader involving cascades
- Tickets fixed:
- [ticket:735]
- [ticket:752]
0.4.0beta3
----------
- SQL types optimization:
- New performance tests show a combined mass-insert/mass-select test as
having 68% fewer function calls than the same test run against 0.3.
- General performance improvement of result set iteration is around 10-20%.
- In types.AbstractType, convert_bind_param() and convert_result_value()
have migrated to callable-returning bind_processor() and
result_processor() methods. If no callable is returned, no pre/post
processing function is called.
- Hooks added throughout base/sql/defaults to optimize the calling of bind
aram/result processors so that method call overhead is minimized.
- Support added for executemany() scenarios such that unneeded "last row id"
logic doesn't kick in, parameters aren't excessively traversed.
- Added 'inherit_foreign_keys' arg to mapper().
- Added support for string date passthrough in sqlite.
- Tickets fixed:
- [ticket:738]
- [ticket:739]
- [ticket:743]
- [ticket:744]
0.4.0beta2
----------
- mssql improvements.
- oracle improvements.
- Auto-commit after LOAD DATA INFILE for mysql.
- A rudimental SessionExtension class has been added, allowing user-defined
functionality to take place at flush(), commit(), and rollback() boundaries.
- Added engine_from_config() function for helping to create_engine() from an
.ini style config.
- base_mapper() becomes a plain attribute.
- session.execute() and scalar() can search for a Table with which to bind from
using the given ClauseElement.
- Session automatically extrapolates tables from mappers with binds, also uses
base_mapper so that inheritance hierarchies bind automatically.
- Moved ClauseVisitor traversal back to inlined non-recursive.
- Tickets fixed:
- [ticket:730]
- [ticket:732]
- [ticket:733]
- [ticket:734]
0.4.0beta1
----------
- orm
- Speed! Along with recent speedups to ResultProxy, total number of function
calls significantly reduced for large loads.
- test/perf/masseagerload.py reports 0.4 as having the fewest number of
function calls across all SA versions (0.1, 0.2, and 0.3).
- New collection_class api and implementation [ticket:213]. Collections are
now instrumented via decorations rather than proxying. You can now have
collections that manage their own membership, and your class instance will
be directly exposed on the relation property. The changes are transparent
for most users.
- InstrumentedList (as it was) is removed, and relation properties no
longer have 'clear()', '.data', or any other added methods beyond those
provided by the collection type. You are free, of course, to add them to
a custom class.
- __setitem__-like assignments now fire remove events for the existing
value, if any.
- dict-likes used as collection classes no longer need to change __iter__
semantics- itervalues() is used by default instead. This is a backwards
incompatible change.
- Subclassing dict for a mapped collection is no longer needed in most
cases. orm.collections provides canned implementations that key objects
by a specified column or a custom function of your choice.
- Collection assignment now requires a compatible type- assigning None to
clear a collection or assigning a list to a dict collection will now
raise an argument error.
- AttributeExtension moved to interfaces, and .delete is now .remove The
event method signature has also been swapped around.
- Major overhaul for Query:
- All selectXXX methods are deprecated. Generative methods are now the
standard way to do things, i.e. filter(), filter_by(), all(), one(),
etc. Deprecated methods are docstring'ed with their new replacements.
- Class-level properties are now usable as query elements... no more
'.c.'! "Class.c.propname" is now superceded by "Class.propname". All
clause operators are supported, as well as higher level operators such
as Class.prop==<some instance> for scalar attributes,
Class.prop.contains(<some instance>) and Class.prop.any(<some
expression>) for collection-based attributes (all are also
negatable). Table-based column expressions as well as columns mounted
on mapped classes via 'c' are of course still fully available and can be
freely mixed with the new attributes. [ticket:643]
- Removed ancient query.select_by_attributename() capability.
- The aliasing logic used by eager loading has been generalized, so that
it also adds full automatic aliasing support to Query. It's no longer
necessary to create an explicit Alias to join to the same tables
multiple times; *even for self-referential relationships*.
- join() and outerjoin() take arguments "aliased=True". Yhis causes
their joins to be built on aliased tables; subsequent calls to
filter() and filter_by() will translate all table expressions (yes,
real expressions using the original mapped Table) to be that of the
Alias for the duration of that join() (i.e. until reset_joinpoint() or
another join() is called).
- join() and outerjoin() take arguments "id=<somestring>". When used
with "aliased=True", the id can be referenced by add_entity(cls,
id=<somestring>) so that you can select the joined instances even if
they're from an alias.
- join() and outerjoin() now work with self-referential relationships!
Using "aliased=True", you can join as many levels deep as desired,
i.e. query.join(['children', 'children'], aliased=True); filter
criterion will be against the rightmost joined table
- Added query.populate_existing(), marks the query to reload all
attributes and collections of all instances touched in the query,
including eagerly-loaded entities. [ticket:660]
- Added eagerload_all(), allows eagerload_all('x.y.z') to specify eager
loading of all properties in the given path.
- Major overhaul for Session:
- New function which "configures" a session called "sessionmaker()". Send
various keyword arguments to this function once, returns a new class
which creates a Session against that stereotype.
- SessionTransaction removed from "public" API. You now can call begin()/
commit()/rollback() on the Session itself.
- Session also supports SAVEPOINT transactions; call begin_nested().
- Session supports two-phase commit behavior when vertically or
horizontally partitioning (i.e., using more than one engine). Use
twophase=True.
- Session flag "transactional=True" produces a session which always places
itself into a transaction when first used. Upon commit(), rollback() or
close(), the transaction ends; but begins again on the next usage.
- Session supports "autoflush=True". This issues a flush() before each
query. Use in conjunction with transactional, and you can just
save()/update() and then query, the new objects will be there. Use
commit() at the end (or flush() if non-transactional) to flush remaining
changes.
- New scoped_session() function replaces SessionContext and assignmapper.
Builds onto "sessionmaker()" concept to produce a class whos Session()
construction returns the thread-local session. Or, call all Session
methods as class methods, i.e. Session.save(foo); Session.commit().
just like the old "objectstore" days.
- Added new "binds" argument to Session to support configuration of
multiple binds with sessionmaker() function.
- A rudimental SessionExtension class has been added, allowing
user-defined functionality to take place at flush(), commit(), and
rollback() boundaries.
- Query-based relation()s available with dynamic_loader(). This is a
*writable* collection (supporting append() and remove()) which is also a
live Query object when accessed for reads. Ideal for dealing with very
large collections where only partial loading is desired.
- flush()-embedded inline INSERT/UPDATE expressions. Assign any SQL
expression, like "sometable.c.column + 1", to an instance's attribute.
Upon flush(), the mapper detects the expression and embeds it directly in
the INSERT or UPDATE statement; the attribute gets deferred on the
instance so it loads the new value the next time you access it.
- A rudimental sharding (horizontal scaling) system is introduced. This
system uses a modified Session which can distribute read and write
operations among multiple databases, based on user-defined functions
defining the "sharding strategy". Instances and their dependents can be
distributed and queried among multiple databases based on attribute
values, round-robin approaches or any other user-defined
system. [ticket:618]
- Eager loading has been enhanced to allow even more joins in more places.
It now functions at any arbitrary depth along self-referential and
cyclical structures. When loading cyclical structures, specify
"join_depth" on relation() indicating how many times you'd like the table
to join to itself; each level gets a distinct table alias. The alias
names themselves are generated at compile time using a simple counting
scheme now and are a lot easier on the eyes, as well as of course
completely deterministic. [ticket:659]
- Added composite column properties. This allows you to create a type which
is represented by more than one column, when using the ORM. Objects of
the new type are fully functional in query expressions, comparisons,
query.get() clauses, etc. and act as though they are regular single-column
scalars... except they're not! Use the function composite(cls, *columns)
inside of the mapper's "properties" dict, and instances of cls will be
created/mapped to a single attribute, comprised of the values correponding
to *columns. [ticket:211]
- Improved support for custom column_property() attributes which feature
correlated subqueries, works better with eager loading now.
- Primary key "collapse" behavior; the mapper will analyze all columns in
its given selectable for primary key "equivalence", that is, columns which
are equivalent via foreign key relationship or via an explicit
inherit_condition. primarily for joined-table inheritance scenarios where
different named PK columns in inheriting tables should "collapse" into a
single-valued (or fewer-valued) primary key. Fixes things like
[ticket:611].
- Joined-table inheritance will now generate the primary key columns of all
inherited classes against the root table of the join only. This implies
that each row in the root table is distinct to a single instance. If for
some rare reason this is not desireable, explicit primary_key settings on
individual mappers will override it.
- When "polymorphic" flags are used with joined-table or single-table
inheritance, all identity keys are generated against the root class of the
inheritance hierarchy; this allows query.get() to work polymorphically
using the same caching semantics as a non-polymorphic get. Note that this
currently does not work with concrete inheritance.
- Secondary inheritance loading: polymorphic mappers can be constructed
*without* a select_table argument. inheriting mappers whose tables were
not represented in the initial load will issue a second SQL query
immediately, once per instance (i.e. not very efficient for large lists),
in order to load the remaining columns.
- Secondary inheritance loading can also move its second query into a
column-level "deferred" load, via the "polymorphic_fetch" argument, which
can be set to 'select' or 'deferred'
- It's now possible to map only a subset of available selectable columns
onto mapper properties, using include_columns/exclude_columns.
[ticket:696].
- Added undefer_group() MapperOption, sets a set of "deferred" columns
joined by a "group" to load as "undeferred".
- Rewrite of the "deterministic alias name" logic to be part of the SQL
layer, produces much simpler alias and label names more in the style of
Hibernate
- sql
- Speed! Clause compilation as well as the mechanics of SQL constructs have
been streamlined and simplified to a signficant degree, for a 20-30%
improvement of the statement construction/compilation overhead of 0.3.
- All "type" keyword arguments, such as those to bindparam(), column(),
Column(), and func.<something>(), renamed to "type_". Those objects still
name their "type" attribute as "type".
- case_sensitive=(True|False) setting removed from schema items, since
checking this state added a lot of method call overhead and there was no
decent reason to ever set it to False. Table and column names which are
all lower case will be treated as case-insenstive (yes we adjust for
Oracle's UPPERCASE style too).
- Transactions:
- Added context manager (with statement) support for transactions.
- Added support for two phase commit, works with mysql and postgres so far.
- Added a subtransaction implementation that uses savepoints.
- Added support for savepoints.
- MetaData:
- Tables can be reflected from the database en-masse without declaring
them in advance. MetaData(engine, reflect=True) will load all tables
present in the database, or use metadata.reflect() for finer control.
- DynamicMetaData has been renamed to ThreadLocalMetaData
- The ThreadLocalMetaData constructor now takes no arguments.
- BoundMetaData has been removed- regular MetaData is equivalent
- Numeric and Float types now have an "asdecimal" flag; defaults to True for
Numeric, False for Float. When True, values are returned as
decimal.Decimal objects; when False, values are returned as float(). The
defaults of True/False are already the behavior for PG and MySQL's DBAPI
modules. [ticket:646]
- New SQL operator implementation which removes all hardcoded operators from
expression structures and moves them into compilation; allows greater
flexibility of operator compilation; for example, "+" compiles to "||"
when used in a string context, or "concat(a,b)" on MySQL; whereas in a
numeric context it compiles to "+". Fixes [ticket:475].
- "Anonymous" alias and label names are now generated at SQL compilation
time in a completely deterministic fashion... no more random hex IDs
- Significant architectural overhaul to SQL elements (ClauseElement). All
elements share a common "mutability" framework which allows a consistent
approach to in-place modifications of elements as well as generative
behavior. Improves stability of the ORM which makes heavy usage of
mutations to SQL expressions.
- select() and union()'s now have "generative" behavior. Methods like
order_by() and group_by() return a *new* instance - the original instance
is left unchanged. Non-generative methods remain as well.
- The internals of select/union vastly simplified- all decision making
regarding "is subquery" and "correlation" pushed to SQL generation phase.
select() elements are now *never* mutated by their enclosing containers or
by any dialect's compilation process [ticket:52] [ticket:569]
- select(scalar=True) argument is deprecated; use select(..).as_scalar().
The resulting object obeys the full "column" interface and plays better
within expressions.
- Added select().with_prefix('foo') allowing any set of keywords to be
placed before the columns clause of the SELECT [ticket:504]
- Added array slice support to row[<index>] [ticket:686]
- Result sets make a better attempt at matching the DBAPI types present in
cursor.description to the TypeEngine objects defined by the dialect, which
are then used for result-processing. Note this only takes effect for
textual SQL; constructed SQL statements always have an explicit type map.
- Result sets from CRUD operations close their underlying cursor immediately
and will also autoclose the connection if defined for the operation; this
allows more efficient usage of connections for successive CRUD operations
with less chance of "dangling connections".
- Column defaults and onupdate Python functions (i.e. passed to
ColumnDefault) may take zero or one arguments; the one argument is the
ExecutionContext, from which you can call "context.parameters[someparam]"
to access the other bind parameter values affixed to the statement
[ticket:559]. The connection used for the execution is available as well
so that you can pre-execute statements.
- Added "explcit" create/drop/execute support for sequences (i.e. you can
pass a "connectable" to each of those methods on Sequence).
- Better quoting of identifiers when manipulating schemas.
- Standardized the behavior for table reflection where types can't be
located; NullType is substituted instead, warning is raised.
- ColumnCollection (i.e. the 'c' attribute on tables) follows dictionary
semantics for "__contains__" [ticket:606]
- engines
- Speed! The mechanics of result processing and bind parameter processing
have been overhauled, streamlined and optimized to issue as little method
calls as possible. Bench tests for mass INSERT and mass rowset iteration
both show 0.4 to be over twice as fast as 0.3, using 68% fewer function
calls.
- You can now hook into the pool lifecycle and run SQL statements or other
logic at new each DBAPI connection, pool check-out and check-in.
- Connections gain a .properties collection, with contents scoped to the
lifetime of the underlying DBAPI connection
- Removed auto_close_cursors and disallow_open_cursors arguments from Pool;
reduces overhead as cursors are normally closed by ResultProxy and
Connection.
- extensions
- proxyengine is temporarily removed, pending an actually working
replacement.
- SelectResults has been replaced by Query. SelectResults /
SelectResultsExt still exist but just return a slightly modified Query
object for backwards-compatibility. join_to() method from SelectResults
isn't present anymore, need to use join().
- mysql
- Table and column names loaded via reflection are now Unicode.
- All standard column types are now supported, including SET.
- Table reflection can now be performed in as little as one round-trip.
- ANSI and ANSI_QUOTES sql modes are now supported.
- Indexes are now reflected.
- postgres
- Added PGArray datatype for using postgres array datatypes.
- oracle
- Very rudimental support for OUT parameters added; use sql.outparam(name,
type) to set up an OUT parameter, just like bindparam(); after execution,
values are avaiable via result.out_parameters dictionary. [ticket:507]
0.3.11
------
- sql
- tweak DISTINCT precedence for clauses like
`func.count(t.c.col.distinct())`
- Fixed detection of internal '$' characters in :bind$params [ticket:719]
- [ticket:768] dont assume join criterion consists only of column objects
- adjusted operator precedence of NOT to match '==' and others, so that
~(x==y) produces NOT (x=y), which is compatible with MySQL < 5.0
(doesn't like "NOT x=y") [ticket:764]
- orm
- added a check for joining from A->B using join(), along two
different m2m tables. this raises an error in 0.3 but is
possible in 0.4 when aliases are used. [ticket:687]
- fixed small exception throw bug in Session.merge()
- fixed bug where mapper, being linked to a join where one table had
no PK columns, would not detect that the joined table had no PK.
- fixed bugs in determining proper sync clauses from custom inherit
conditions [ticket:769]
- backref remove object operation doesn't fail if the other-side
collection doesn't contain the item, supports noload collections
[ticket:813]
- engine
- fixed another occasional race condition which could occur
when using pool with threadlocal setting
- mysql
- fixed specification of YEAR columns when generating schema
- mssql
- added support for TIME columns (simulated using DATETIME) [ticket:679]
- added support for BIGINT, MONEY, SMALLMONEY, UNIQUEIDENTIFIER and
SQL_VARIANT [ticket:721]
- index names are now quoted when dropping from reflected tables
[ticket:684]
- can now specify a DSN for PyODBC, using a URI like mssql:///?dsn=bob
- postgres
- when reflecting tables from alternate schemas, the "default" placed upon
the primary key, i.e. usually a sequence name, has the "schema" name
unconditionally quoted, so that schema names which need quoting are fine.
its slightly unnecessary for schema names which don't need quoting
but not harmful.
- sqlite
- passthrough for stringified dates
- firebird
- supports_sane_rowcount() set to False due to ticket #370 (right way).
- fixed reflection of Column's nullable property.
- oracle
- removed LONG_STRING, LONG_BINARY from "binary" types, so type objects
don't try to read their values as LOB [ticket:622], [ticket:751]
0.3.10
- general
- a new mutex that was added in 0.3.9 causes the pool_timeout
feature to fail during a race condition; threads would
raise TimeoutError immediately with no delay if many threads
push the pool into overflow at the same time. this issue has been
fixed.
- sql
- got connection-bound metadata to work with implicit execution
- foreign key specs can have any chararcter in their identifiers
[ticket:667]
- added commutativity-awareness to binary clause comparisons to
each other, improves ORM lazy load optimization [ticket:664]
- orm
- cleanup to connection-bound sessions, SessionTransaction
- postgres
- fixed max identifier length (63) [ticket:571]
0.3.9
- general
- better error message for NoSuchColumnError [ticket:607]
- finally figured out how to get setuptools version in, available
as sqlalchemy.__version__ [ticket:428]
- the various "engine" arguments, such as "engine", "connectable",
"engine_or_url", "bind_to", etc. are all present, but deprecated.
they all get replaced by the single term "bind". you also
set the "bind" of MetaData using
metadata.bind = <engine or connection>
- ext
- iteration over dict association proxies is now dict-like, not
InstrumentedList-like (e.g. over keys instead of values)
- association proxies no longer bind tightly to source collections
[ticket:597], and are constructed with a thunk instead
- added selectone_by() to assignmapper
- orm
- forwards-compatibility with 0.4: added one(), first(), and
all() to Query. almost all Query functionality from 0.4 is
present in 0.3.9 for forwards-compat purposes.
- reset_joinpoint() really really works this time, promise ! lets
you re-join from the root:
query.join(['a', 'b']).filter(<crit>).reset_joinpoint().\
join(['a', 'c']).filter(<some other crit>).all()
in 0.4 all join() calls start from the "root"
- added synchronization to the mapper() construction step, to avoid
thread collisions when pre-existing mappers are compiling in a
different thread [ticket:613]
- a warning is issued by Mapper when two primary key columns of the
same name are munged into a single attribute. this happens frequently
when mapping to joins (or inheritance).
- synonym() properties are fully supported by all Query joining/
with_parent operations [ticket:598]
- fixed very stupid bug when deleting items with many-to-many
uselist=False relations
- remember all that stuff about polymorphic_union ? for
joined table inheritance ? Funny thing...
You sort of don't need it for joined table inheritance, you
can just string all the tables together via outerjoin().
The UNION still applies if concrete tables are involved,
though (since nothing to join them on).
- small fix to eager loading to better work with eager loads
to polymorphic mappers that are using a straight "outerjoin"
clause
- sql
- ForeignKey to a table in a schema thats not the default schema
requires the schema to be explicit; i.e. ForeignKey('alt_schema.users.id')
- MetaData can now be constructed with an engine or url as the first
argument, just like BoundMetaData
- BoundMetaData is now deprecated, and MetaData is a direct substitute.
- DynamicMetaData has been renamed to ThreadLocalMetaData. the
DynamicMetaData name is deprecated and is an alias for ThreadLocalMetaData
or a regular MetaData if threadlocal=False
- composite primary key is represented as a non-keyed set to allow for
composite keys consisting of cols with the same name; occurs within a
Join. helps inheritance scenarios formulate correct PK.
- improved ability to get the "correct" and most minimal set of primary key
columns from a join, equating foreign keys and otherwise equated columns.
this is also mostly to help inheritance scenarios formulate the best
choice of primary key columns. [ticket:185]
- added 'bind' argument to Sequence.create()/drop(), ColumnDefault.execute()
- columns can be overridden in a reflected table with a "key"
attribute different than the column's name, including for primary key
columns [ticket:650]
- fixed "ambiguous column" result detection, when dupe col names exist
in a result [ticket:657]
- some enhancements to "column targeting", the ability to match a column
to a "corresponding" column in another selectable. this affects mostly
ORM ability to map to complex joins
- MetaData and all SchemaItems are safe to use with pickle. slow
table reflections can be dumped into a pickled file to be reused later.
Just reconnect the engine to the metadata after unpickling. [ticket:619]
- added a mutex to QueuePool's "overflow" calculation to prevent a race
condition that can bypass max_overflow
- fixed grouping of compound selects to give correct results. will break
on sqlite in some cases, but those cases were producing incorrect
results anyway, sqlite doesn't support grouped compound selects
[ticket:623]
- fixed precedence of operators so that parenthesis are correctly applied
[ticket:620]
- calling <column>.in_() (i.e. with no arguments) will return
"CASE WHEN (<column> IS NULL) THEN NULL ELSE 0 END = 1)", so that
NULL or False is returned in all cases, rather than throwing an error
[ticket:545]
- fixed "where"/"from" criterion of select() to accept a unicode string
in addition to regular string - both convert to text()
- added standalone distinct() function in addition to column.distinct()
[ticket:558]
- result.last_inserted_ids() should return a list that is identically
sized to the primary key constraint of the table. values that were
"passively" created and not available via cursor.lastrowid will be None.
- long-identifier detection fixed to use > rather than >= for
max ident length [ticket:589]
- fixed bug where selectable.corresponding_column(selectable.c.col)
would not return selectable.c.col, if the selectable is a join
of a table and another join involving the same table. messed
up ORM decision making [ticket:593]
- added Interval type to types.py [ticket:595]
- mysql
- fixed catching of some errors that imply a dropped connection [ticket:625]
- fixed escaping of the modulo operator [ticket:624]
- added 'fields' to reserved words [ticket:590]
- various reflection enhancement/fixes
- oracle
- datetime fixes: got subsecond TIMESTAMP to work [ticket:604],
added OracleDate which supports types.Date with only year/month/day
- added dialect flag "auto_convert_lobs", defaults to True; will cause any
LOB objects detected in a result set to be forced into OracleBinary
so that the LOB is read() automatically, if no typemap was present
(i.e., if a textual execute() was issued).
- mod operator '%' produces MOD [ticket:624]
- converts cx_oracle datetime objects to Python datetime.datetime when
Python 2.3 used [ticket:542]
- fixed unicode conversion in Oracle TEXT type
- postgres
- fixed escaping of the modulo operator [ticket:624]
- added support for reflection of domains [ticket:570]
- types which are missing during reflection resolve to Null type
instead of raising an error
- the fix in "schema" above fixes reflection of foreign keys from an
alt-schema table to a public schema table
- sqlite
- rearranged dialect initialization so it has time to warn about pysqlite1
being too old.
- sqlite better handles datetime/date/time objects mixed and matched
with various Date/Time/DateTime columns
- string PK column inserts dont get overwritten with OID [ticket:603]
- mssql
- fix port option handling for pyodbc [ticket:634]
- now able to reflect start and increment values for identity columns
- preliminary support for using scope_identity() with pyodbc
0.3.8
- engines
- added detach() to Connection, allows underlying DBAPI connection
to be detached from its pool, closing on dereference/close()
instead of being reused by the pool.
- added invalidate() to Connection, immediately invalidates the
Connection and its underlying DBAPI connection.
- sql
- _Label class overrides compare_self to return its ultimate
object. meaning, if you say someexpr.label('foo') == 5, it
produces the correct "someexpr == 5".
- _Label propagates "_hide_froms()" so that scalar selects
behave more properly with regards to FROM clause #574
- fix to long name generation when using oid_column as an order by
(oids used heavily in mapper queries)
- significant speed improvement to ResultProxy, pre-caches
TypeEngine dialect implementations and saves on function calls
per column
- parenthesis are applied to clauses via a new _Grouping
construct. uses operator precedence to more intelligently apply
parenthesis to clauses, provides cleaner nesting of clauses
(doesnt mutate clauses placed in other clauses, i.e. no 'parens'
flag)
- added 'modifier' keyword, works like func.<foo> except does not
add parenthesis. e.g. select([modifier.DISTINCT(...)]) etc.
- removed "no group by's in a select thats part of a UNION"
restriction [ticket:578]
- orm
- added reset_joinpoint() method to Query, moves the "join point"
back to the starting mapper. 0.4 will change the behavior of
join() to reset the "join point" in all cases so this is an
interim method. for forwards compatibility, ensure joins across
multiple relations are specified using a single join(), i.e.
join(['a', 'b', 'c']).
- fixed bug in query.instances() that wouldnt handle more than
on additional mapper or one additional column.
- "delete-orphan" no longer implies "delete". ongoing effort to
separate the behavior of these two operations.
- many-to-many relationships properly set the type of bind params
for delete operations on the association table
- many-to-many relationships check that the number of rows deleted
from the association table by a delete operation matches the
expected results
- session.get() and session.load() propagate **kwargs through to
query
- fix to polymorphic query which allows the original
polymorphic_union to be embedded into a correlated subquery
[ticket:577]
- fix to select_by(<propname>=<object instance>) -style joins in
conjunction with many-to-many relationships, bug introduced in
r2556
- the "primary_key" argument to mapper() is propagated to the
"polymorphic" mapper. primary key columns in this list get
normalized to that of the mapper's local table.
- restored logging of "lazy loading clause" under
sa.orm.strategies logger, got removed in 0.3.7
- improved support for eagerloading of properties off of mappers
that are mapped to select() statements; i.e. eagerloader is
better at locating the correct selectable with which to attach
its LEFT OUTER JOIN.
- mysql
- Nearly all MySQL column types are now supported for declaration
and reflection. Added NCHAR, NVARCHAR, VARBINARY, TINYBLOB,
LONGBLOB, YEAR
- The sqltypes.Binary passthrough now always builds a BLOB,
avoiding problems with very old database versions
- support for column-level CHARACTER SET and COLLATE declarations,
as well as ASCII, UNICODE, NATIONAL and BINARY shorthand.
- firebird
- set max identifier length to 31
- supports_sane_rowcount() set to False due to ticket #370.
versioned_id_col feature wont work in FB.
- some execution fixes
-extensions
- new association proxy implementation, implementing complete
proxies to list, dict and set-based relation collections
- added orderinglist, a custom list class that synchronizes an
object attribute with that object's position in the list
- small fix to SelectResultsExt to not bypass itself during
select().
- added filter(), filter_by() to assignmapper
0.3.7
- engines
- warnings module used for issuing warnings (instead of logging)
- cleanup of DBAPI import strategies across all engines
[ticket:480]
- refactoring of engine internals which reduces complexity,
number of codepaths; places more state inside of ExecutionContext
to allow more dialect control of cursor handling, result sets.
ResultProxy totally refactored and also has two versions of
"buffered" result sets used for different purposes.
- server side cursor support fully functional in postgres
[ticket:514].
- improved framework for auto-invalidation of connections that have
lost their underlying database, via dialect-specific detection
of exceptions corresponding to that database's disconnect
related error messages. Additionally, when a "connection no
longer open" condition is detected, the entire connection pool
is discarded and replaced with a new instance. #516
- the dialects within sqlalchemy.databases become a setuptools
entry points. loading the built-in database dialects works the
same as always, but if none found will fall back to trying
pkg_resources to load an external module [ticket:521]
- Engine contains a "url" attribute referencing the url.URL object
used by create_engine().
- sql:
- keys() of result set columns are not lowercased, come back
exactly as they're expressed in cursor.description. note this
causes colnames to be all caps in oracle.
- preliminary support for unicode table names, column names and
SQL statements added, for databases which can support them.
Works with sqlite and postgres so far. Mysql *mostly* works
except the has_table() function does not work. Reflection
works too.
- the Unicode type is now a direct subclass of String, which now
contains all the "convert_unicode" logic. This helps the variety
of unicode situations that occur in db's such as MS-SQL to be
better handled and allows subclassing of the Unicode datatype.
[ticket:522]
- ClauseElements can be used in in_() clauses now, such as bind
parameters, etc. #476
- reverse operators implemented for `CompareMixin` elements,
allows expressions like "5 + somecolumn" etc. #474
- the "where" criterion of an update() and delete() now correlates
embedded select() statements against the table being updated or
deleted. this works the same as nested select() statement
correlation, and can be disabled via the correlate=False flag on
the embedded select().
- column labels are now generated in the compilation phase, which
means their lengths are dialect-dependent. So on oracle a label
that gets truncated to 30 chars will go out to 63 characters
on postgres. Also, the true labelname is always attached as the
accessor on the parent Selectable so theres no need to be aware
of the "truncated" label names [ticket:512].
- column label and bind param "truncation" also generate
deterministic names now, based on their ordering within the
full statement being compiled. this means the same statement
will produce the same string across application restarts and
allowing DB query plan caching to work better.
- the "mini" column labels generated when using subqueries, which
are to work around glitchy SQLite behavior that doesnt understand
"foo.id" as equivalent to "id", are now only generated in the case
that those named columns are selected from (part of [ticket:513])
- the label() method on ColumnElement will properly propagate the
TypeEngine of the base element out to the label, including a label()
created from a scalar=True select() statement.
- MS-SQL better detects when a query is a subquery and knows not to
generate ORDER BY phrases for those [ticket:513]
- fix for fetchmany() "size" argument being positional in most
dbapis [ticket:505]
- sending None as an argument to func.<something> will produce
an argument of NULL
- query strings in unicode URLs get keys encoded to ascii
for **kwargs compat
- slight tweak to raw execute() change to also support tuples
for positional parameters, not just lists [ticket:523]
- fix to case() construct to propagate the type of the first
WHEN condition as the return type of the case statement
- orm:
- fixed critical issue when, after options(eagerload()) is used,
the mapper would then always apply query "wrapping" behavior
for all subsequent LIMIT/OFFSET/DISTINCT queries, even if no
eager loading was applied on those subsequent queries.
- added query.with_parent(someinstance) method. searches for
target instance using lazy join criterion from parent instance.
takes optional string "property" to isolate the desired relation.
also adds static Query.query_from_parent(instance, property)
version. [ticket:541]
- improved query.XXX_by(someprop=someinstance) querying to use
similar methodology to with_parent, i.e. using the "lazy" clause
which prevents adding the remote instance's table to the SQL,
thereby making more complex conditions possible [ticket:554]
- added generative versions of aggregates, i.e. sum(), avg(), etc.
to query. used via query.apply_max(), apply_sum(), etc.
#552
- fix to using distinct() or distinct=True in combination with
join() and similar
- corresponding to label/bindparam name generation, eager loaders
generate deterministic names for the aliases they create using
md5 hashes.
- improved/fixed custom collection classes when giving it "set"/
"sets.Set" classes or subclasses (was still looking for append()
methods on them during lazy loads)
- restored old "column_property()" ORM function (used to be called
"column()") to force any column expression to be added as a property
on a mapper, particularly those that aren't present in the mapped
selectable. this allows "scalar expressions" of any kind to be
added as relations (though they have issues with eager loads).
- fix to many-to-many relationships targeting polymorphic mappers
[ticket:533]
- making progress with session.merge() as well as combining its
usage with entity_name [ticket:543]
- the usual adjustments to relationships between inheriting mappers,
in this case establishing relation()s to subclass mappers where
the join conditions come from the superclass' table
- informix:
- informix support added ! courtesy James Zhang, who put a ton
of effort in.
- sqlite:
- removed silly behavior where sqlite would reflect UNIQUE indexes
as part of the primary key (?!)
- oracle:
- small fix to allow successive compiles of the same SELECT object
which features LIMIT/OFFSET. oracle dialect needs to modify
the object to have ROW_NUMBER OVER and wasn't performing
the full series of steps on successive compiles.
- mysql
- support for SSL arguments given as inline within URL query string,
prefixed with "ssl_", courtesy terjeros@gmail.com.
- mysql uses "DESCRIBE [<schemaname>].<tablename>", catching exceptions
if table doesnt exist, in order to determine if a table exists.
this supports unicode table names as well as schema names. tested
with MySQL5 but should work with 4.1 series as well. (#557)
- extensions
- big fix to AssociationProxy so that multiple AssociationProxy
objects can be associated with a single association collection.
- assign_mapper names methods according to their keys (i.e. __name__)
#551
- mssql
- pyodbc is now the preferred DB-API for MSSQL, and if no module is
specifically requested, will be loaded first on a module probe.
- The @@SCOPE_IDENTITY is now used instead of @@IDENTITY. This
behavior may be overridden with the engine_connect
"use_scope_identity" keyword parameter, which may also be specified
in the dburi.
0.3.6
- sql:
- bindparam() names are now repeatable! specify two
distinct bindparam()s with the same name in a single statement,
and the key will be shared. proper positional/named args translate
at compile time. for the old behavior of "aliasing" bind parameters
with conflicting names, specify "unique=True" - this option is
still used internally for all the auto-genererated (value-based)
bind parameters.
- slightly better support for bind params as column clauses, either
via bindparam() or via literal(), i.e. select([literal('foo')])
- MetaData can bind to an engine either via "url" or "engine" kwargs
to constructor, or by using connect() method. BoundMetaData is
identical to MetaData except engine_or_url param is required.
DynamicMetaData is the same and provides thread-local connections be
default.
- exists() becomes useable as a standalone selectable, not just in a
WHERE clause, i.e. exists([columns], criterion).select()
- correlated subqueries work inside of ORDER BY, GROUP BY
- fixed function execution with explicit connections, i.e.
conn.execute(func.dosomething())
- use_labels flag on select() wont auto-create labels for literal text
column elements, since we can make no assumptions about the text. to
create labels for literal columns, you can say "somecol AS
somelabel", or use literal_column("somecol").label("somelabel")
- quoting wont occur for literal columns when they are "proxied" into
the column collection for their selectable (is_literal flag is
propagated). literal columns are specified via
literal_column("somestring").
- added "fold_equivalents" boolean argument to Join.select(), which
removes 'duplicate' columns from the resulting column clause that
are known to be equivalent based on the join condition. this is of
great usage when constructing subqueries of joins which Postgres
complains about if duplicate column names are present.
- fixed use_alter flag on ForeignKeyConstraint [ticket:503]
- fixed usage of 2.4-only "reversed" in topological.py [ticket:506]
- for hackers, refactored the "visitor" system of ClauseElement and
SchemaItem so that the traversal of items is controlled by the
ClauseVisitor itself, using the method visitor.traverse(item).
accept_visitor() methods can still be called directly but will not
do any traversal of child items. ClauseElement/SchemaItem now have a
configurable get_children() method to return the collection of child
elements for each parent object. This allows the full traversal of
items to be clear and unambiguous (as well as loggable), with an
easy method of limiting a traversal (just pass flags which are
picked up by appropriate get_children() methods). [ticket:501]
- the "else_" parameter to the case statement now properly works when
set to zero.
- orm:
- the full featureset of the SelectResults extension has been merged
into a new set of methods available off of Query. These methods
all provide "generative" behavior, whereby the Query is copied
and a new one returned with additional criterion added.
The new methods include:
filter() - applies select criterion to the query
filter_by() - applies "by"-style criterion to the query
avg() - return the avg() function on the given column
join() - join to a property (or across a list of properties)
outerjoin() - like join() but uses LEFT OUTER JOIN
limit()/offset() - apply LIMIT/OFFSET
range-based access which applies limit/offset:
session.query(Foo)[3:5]
distinct() - apply DISTINCT
list() - evaluate the criterion and return results
no incompatible changes have been made to Query's API and no methods
have been deprecated. Existing methods like select(), select_by(),
get(), get_by() all execute the query at once and return results
like they always did. join_to()/join_via() are still there although
the generative join()/outerjoin() methods are easier to use.
- the return value for multiple mappers used with instances() now
returns a cartesian product of the requested list of mappers,
represented as a list of tuples. this corresponds to the documented
behavior. So that instances match up properly, the "uniquing" is
disabled when this feature is used.
- Query has add_entity() and add_column() generative methods. these
will add the given mapper/class or ColumnElement to the query at
compile time, and apply them to the instances() method. the user is
responsible for constructing reasonable join conditions (otherwise
you can get full cartesian products). result set is the list of
tuples, non-uniqued.
- strings and columns can also be sent to the *args of instances()
where those exact result columns will be part of the result tuples.
- a full select() construct can be passed to query.select() (which
worked anyway), but also query.selectfirst(), query.selectone()
which will be used as is (i.e. no query is compiled). works
similarly to sending the results to instances().
- eager loading will not "aliasize" "order by" clauses that were
placed in the select statement by something other than the eager
loader itself, to fix possibility of dupe columns as illustrated in
[ticket:495]. however, this means you have to be more careful with
the columns placed in the "order by" of Query.select(), that you
have explicitly named them in your criterion (i.e. you cant rely on
the eager loader adding them in for you)
- added a handy multi-use "identity_key()" method to Session, allowing
the generation of identity keys for primary key values, instances,
and rows, courtesy Daniel Miller
- many-to-many table will be properly handled even for operations that
occur on the "backref" side of the operation [ticket:249]
- added "refresh-expire" cascade [ticket:492]. allows refresh() and
expire() calls to propagate along relationships.
- more fixes to polymorphic relations, involving proper lazy-clause
generation on many-to-one relationships to polymorphic mappers
[ticket:493]. also fixes to detection of "direction", more specific
targeting of columns that belong to the polymorphic union vs. those
that dont.
- some fixes to relationship calcs when using "viewonly=True" to pull
in other tables into the join condition which arent parent of the
relationship's parent/child mappings
- flush fixes on cyclical-referential relationships that contain
references to other instances outside of the cyclical chain, when
some of the objects in the cycle are not actually part of the flush
- put an aggressive check for "flushing object A with a collection of
B's, but you put a C in the collection" error condition - **even if
C is a subclass of B**, unless B's mapper loads polymorphically.
Otherwise, the collection will later load a "B" which should be a
"C" (since its not polymorphic) which breaks in bi-directional
relationships (i.e. C has its A, but A's backref will lazyload it as
a different instance of type "B") [ticket:500] This check is going
to bite some of you who do this without issues, so the error message
will also document a flag "enable_typechecks=False" to disable this
checking. But be aware that bi-directional relationships in
particular become fragile without this check.
- extensions:
- options() method on SelectResults now implemented "generatively"
like the rest of the SelectResults methods [ticket:472]. But
you're going to just use Query now anyway.
- query() method is added by assignmapper. this helps with
navigating to all the new generative methods on Query.
- ms-sql:
- removed seconds input on DATE column types (probably
should remove the time altogether)
- null values in float fields no longer raise errors
- LIMIT with OFFSET now raises an error (MS-SQL has no OFFSET support)
- added an facility to use the MSSQL type VARCHAR(max) instead of TEXT
for large unsized string fields. Use the new "text_as_varchar" to
turn it on. [ticket:509]
- ORDER BY clauses without a LIMIT are now stripped in subqueries, as
MS-SQL forbids this usage
- cleanup of module importing code; specifiable DB-API module; more
explicit ordering of module preferences. [ticket:480]
- oracle:
- got binary working for any size input ! cx_oracle works fine,
it was my fault as BINARY was being passed and not BLOB for
setinputsizes (also unit tests werent even setting input sizes).
- also fixed CLOB read/write on a separate changeset.
- auto_setinputsizes defaults to True for Oracle, fixed cases where
it improperly propagated bad types.
- mysql:
- added a catchall **kwargs to MSString, to help reflection of
obscure types (like "varchar() binary" in MS 4.0)
- added explicit MSTimeStamp type which takes effect when using
types.TIMESTAMP.
0.3.5
- sql:
- the value of "case_sensitive" defaults to True now, regardless of the
casing of the identifier, unless specifically set to False. this is
because the object might be label'ed as something else which does
contain mixed case, and propigating "case_sensitive=False" breaks that.
Other fixes to quoting when using labels and "fake" column objects
- added a "supports_execution()" method to ClauseElement, so that
individual kinds of clauses can express if they are appropriate for
executing...such as, you can execute a "select", but not a "Table" or a
"Join".
- fixed argument passing to straight textual execute() on engine,
connection. can handle *args or a list instance for positional, **kwargs
or a dict instance for named args, or a list of list or dicts to invoke
executemany()
- small fix to BoundMetaData to accept unicode or string URLs
- fixed named PrimaryKeyConstraint generation [ticket:466] courtesy
andrija at gmail
- fixed generation of CHECK constraints on columns [ticket:464]
- fixes to tometadata() operation to propagate Constraints at column and
table level
- oracle:
- when returning "rowid" as the ORDER BY column or in use with ROW_NUMBER
OVER, oracle dialect checks the selectable its being applied to and will
switch to table PK if not applicable, i.e. for a UNION. checking for
DISTINCT, GROUP BY (other places that rowid is invalid) still a TODO.
allows polymorphic mappings to function, [ticket:436]
- sequences on a non-pk column will properly fire off on INSERT
- added PrefetchingResultProxy support to pre-fetch LOB columns when they
are known to be present, fixes [ticket:435]
- implemented reflection of tables based on synonyms, including across
dblinks [ticket:379]
- issues a log warning when a related table cant be reflected due to
certain permission errors [ticket:363]
- mysql:
- fix to reflection on older DB's that might return array() type for
"show variables like" statements
- postgres:
- better reflection of sequences for alternate-schema Tables [ticket:442]
- sequences on a non-pk column will properly fire off on INSERT
- added PGInterval type [ticket:460], PGInet type [ticket:444]
- mssql:
- preliminary support for pyodbc (Yay!) [ticket:419]
- better support for NVARCHAR types added [ticket:298]
- fix for commit logic on pymssql
- fix for query.get() with schema [ticket:456]
- fix for non-integer relationships [ticket:473]
- DB-API module now selectable at run-time [ticket:419]
- now passes many more unit tests [tickets:422, 481, 415]
- better unittest compatibility with ANSI functions [ticket:479]
- improved support for implicit sequence PK columns with auto-insert
[ticket:415]
- fix for blank password in adodbapi [ticket:371]
- fixes to get unit tests working with pyodbc [ticket:481]
- fix to auto_identity_insert on db-url query
- added query_timeout to db-url query parms. currently works only for
pymssql
- tested with pymssql 0.8.0 (which is now LGPL)
- orm bugs:
- another refactoring to relationship calculation. Allows more accurate
ORM behavior with relationships from/to/between mappers, particularly
polymorphic mappers, also their usage with Query, SelectResults. tickets
include [ticket:439], [ticket:441], [ticket:448].
- removed deprecated method of specifying custom collections on classes;
you must now use the "collection_class" option. the old way was
beginning to produce conflicts when people used assign_mapper(), which
now patches an "options" method, in conjunction with a relationship
named "options". (relationships take precedence over monkeypatched
assign_mapper methods).
- extension() query option propagates to Mapper._instance() method so that
all loading-related methods get called [ticket:454]
- eager relation to an inheriting mapper wont fail if no rows returned for
the relationship.
- eager relation loading bug fixed for eager relation on multiple
descendant classes [ticket:486]
- fix for very large topological sorts, courtesy ants.aasma at gmail
[ticket:423]
- eager loading is slightly more strict about detecting "self-referential"
relationships, specifically between polymorphic mappers. this results in
an "eager degrade" to lazy loading.
- improved support for complex queries embedded into "where" criterion for
query.select() [ticket:449]
- mapper options like eagerload(), lazyload(), deferred(), will work for
"synonym()" relationships [ticket:485]
- fixed bug where cascade operations incorrectly included deleted
collection items in the cascade [ticket:445]
- fixed relationship deletion error when one-to-many child item is moved
to a new parent in a single unit of work [ticket:478]
- fixed relationship deletion error where parent/child with a single
column as PK/FK on the child would raise a "blank out the primary key"
error, if manually deleted or "delete" cascade without "delete-orphan"
was used
- fix to deferred so that load operation doesnt mistakenly occur when only
PK col attributes are set
- orm enhancements:
- implemented foreign_keys argument to mapper [ticket:385]. use in
conjunction with primaryjoin/secondaryjoin arguments to specify/override
foreign keys defined on the Table instance.
- contains_eager('foo') automatically implies eagerload('foo')
- added "alias" argument to contains_eager(). use it to specify the string
name or Alias instance of an alias used in the query for the eagerly
loaded child items. easier to use than "decorator"
- added "contains_alias()" option for result set mapping to an alias of
the mapped table
- added support for py2.5 "with" statement with SessionTransaction
[ticket:468]
- extensions:
- added distinct() method to SelectResults. generally should only make a
difference when using count().
- added options() method to SelectResults, equivalent to query.options()
[ticket:472]
- added optional __table_opts__ dictionary to ActiveMapper, will send kw
options to Table objects [ticket:462]
- added selectfirst(), selectfirst_by() to assign_mapper [ticket:467]
0.3.4
- general:
- global "insure"->"ensure" change. in US english "insure" is actually
largely interchangeable with "ensure" (so says the dictionary), so I'm not
completely illiterate, but its definitely sub-optimal to "ensure" which is
non-ambiguous.
- sql:
- added "fetchmany()" support to ResultProxy
- added support for column "key" attribute to be useable in
row[<key>]/row.<key>
- changed "BooleanExpression" to subclass from "BinaryExpression", so that
boolean expressions can also follow column-clause behaviors (i.e. label(),
etc).
- trailing underscores are trimmed from func.<xxx> calls, such as func.if_()
- fix to correlation of subqueries when the column list of the select
statement is constructed with individual calls to append_column(); this
fixes an ORM bug whereby nested select statements were not getting
correlated with the main select generated by the Query object.
- another fix to subquery correlation so that a subquery which has only one
FROM element will *not* correlate that single element, since at least one
FROM element is required in a query.
- default "timezone" setting is now False. this corresponds to Python's
datetime behavior as well as Postgres' timestamp/time types (which is the
only timezone-sensitive dialect at the moment) [ticket:414]
- the "op()" function is now treated as an "operation", rather than a
"comparison". the difference is, an operation produces a BinaryExpression
from which further operations can occur whereas comparison produces the
more restrictive BooleanExpression
- trying to redefine a reflected primary key column as non-primary key raises
an error
- type system slightly modified to support TypeDecorators that can be
overridden by the dialect (ok, thats not very clear, it allows the mssql
tweak below to be possible)
- mssql:
- added an NVarchar type (produces NVARCHAR), also MSUnicode which provides
Unicode-translation for the NVarchar regardless of dialect convert_unicode
setting.
- postgres:
- fix to the initial checkfirst for tables to take current schema into
account [ticket:424]
- postgres has an optional "server_side_cursors=True" flag which will utilize
server side cursors. these are appropriate for fetching only partial
results and are necessary for working with very large unbounded result
sets. While we'd like this to be the default behavior, different
environments seem to have different results and the causes have not been
isolated so we are leaving the feature off by default for now. Uses an
apparently undocumented psycopg2 behavior recently discovered on the
psycopg mailing list.
- added "BIGSERIAL" support for postgres table with
PGBigInteger/autoincrement
- fixes to postgres reflection to better handle when schema names are
present; thanks to jason (at) ncsmags.com [ticket:402]
- mysql:
- mysql is inconsistent with what kinds of quotes it uses in foreign keys
during a SHOW CREATE TABLE, reflection updated to accomodate for all three
styles [ticket:420]
- mysql table create options work on a generic passthru now, i.e. Table(...,
mysql_engine='InnoDB', mysql_collate="latin1_german2_ci",
mysql_auto_increment="5", mysql_<somearg>...), helps [ticket:418]
- firebird:
- order of constraint creation puts primary key first before all other
constraints; required for firebird, not a bad idea for others [ticket:408]
- Firebird fix to autoload multifield foreign keys [ticket:409]
- Firebird NUMERIC type properly handles a type without precision
[ticket:409]
- oracle:
- *slight* support for binary, but still need to figure out how to insert
reasonably large values (over 4K). requires auto_setinputsizes=True sent to
create_engine(), rows must be fully fetched individually, etc.
- orm:
- poked the first hole in the can of worms: saying
query.select_by(somerelationname=someinstance) will create the join of the
primary key columns represented by "somerelationname"'s mapper to the
actual primary key in "someinstance".
- reworked how relations interact with "polymorphic" mappers, i.e. mappers
that have a select_table as well as polymorphic flags. better determination
of proper join conditions, interaction with user- defined join conditions,
and support for self-referential polymorphic mappers.
- related to polymorphic mapping relations, some deeper error checking when
compiling relations, to detect an ambiguous "primaryjoin" in the case that
both sides of the relationship have foreign key references in the primary
join condition. also tightened down conditions used to locate "relation
direction", associating the "foreignkey" of the relationship with the
"primaryjoin"
- a little bit of improvement to the concept of a "concrete" inheritance
mapping, though that concept is not well fleshed out yet (added test case
to support concrete mappers on top of a polymorphic base).
- fix to "proxy=True" behavior on synonym()
- fixed bug where delete-orphan basically didn't work with many-to-many
relationships [ticket:427], backref presence generally hid the symptom
- added a mutex to the mapper compilation step. ive been reluctant to add any
kind of threading anything to SA but this is one spot that its its really
needed since mappers are typically "global", and while their state does not
change during normal operation, the initial compilation step does modify
internal state significantly, and this step usually occurs not at
module-level initialization time (unless you call compile()) but at
first-request time
- basic idea of "session.merge()" actually implemented. needs more testing.
- added "compile_mappers()" function as a shortcut to compiling all mappers
- fix to MapperExtension create_instance so that entity_name properly
associated with new instance
- speed enhancements to ORM object instantiation, eager loading of rows
- invalid options sent to 'cascade' string will raise an exception
[ticket:406]
- fixed bug in mapper refresh/expire whereby eager loaders didnt properly
re-populate item lists [ticket:407]
- fix to post_update to ensure rows are updated even for non insert/delete
scenarios [ticket:413]
- added an error message if you actually try to modify primary key values on
an entity and then flush it [ticket:412]
- extensions
- added "validate=False" argument to assign_mapper, if True will ensure that
only mapped attributes are named [ticket:426]
- assign_mapper gets "options", "instances" functions added (i.e.
MyClass.instances())
0.3.3
- string-based FROM clauses fixed, i.e. select(..., from_obj=["sometext"])
- fixes to passive_deletes flag, lazy=None (noload) flag
- added example/docs for dealing with large collections
- added object_session() method to sqlalchemy namespace
- fixed QueuePool bug whereby its better able to reconnect to a database
that was not reachable (thanks to Sébastien Lelong), also fixed dispose()
method
- patch that makes MySQL rowcount work correctly! [ticket:396]
- fix to MySQL catch of 2006/2014 errors to properly re-raise OperationalError
exception
0.3.2
- major connection pool bug fixed. fixes MySQL out of sync
errors, will also prevent transactions getting rolled back
accidentally in all DBs [ticket:387]
- major speed enhancements vs. 0.3.1, to bring speed
back to 0.2.8 levels
- made conditional dozens of debug log calls that were
time-intensive to generate log messages
- fixed bug in cascade rules whereby the entire object graph
could be unnecessarily cascaded on the save/update cascade
- various speedups in attributes module
- identity map in Session is by default *no longer weak referencing*.
to have it be weak referencing, use create_session(weak_identity_map=True)
fixes [ticket:388]
- MySQL detects errors 2006 (server has gone away) and 2014
(commands out of sync) and invalidates the connection on which it occured.
- MySQL bool type fix: [ticket:307]
- postgres reflection fixes: [ticket:349] [ticket:382]
- added keywords for EXCEPT, INTERSECT, EXCEPT ALL, INTERSECT ALL
[ticket:247]
- assign_mapper in assignmapper extension returns the created mapper
[changeset:2110]
- added label() function to Select class, when scalar=True is used
to create a scalar subquery
i.e. "select x, y, (select max(foo) from table) AS foomax from table"
- added onupdate and ondelete keyword arguments to ForeignKey; propagate
to underlying ForeignKeyConstraint if present. (dont propagate in the
other direction, however)
- fix to session.update() to preserve "dirty" status of incoming object
- sending a selectable to an IN via the in_() function no longer creates
a "union" out of multiple selects; only one selectable to a the in_() function
is allowed now (make a union yourself if union is needed)
- improved support for disabling save-update cascade via cascade="none" etc.
- added "remote_side" argument to relation(), used only with self-referential
mappers to force the direction of the parent/child relationship. replaces
the usage of the "foreignkey" parameter for "switching" the direction.
"foreignkey" argument is deprecated for all uses and will eventually
be replaced by an argument dedicated to ForeignKey specification on mappers.
0.3.1
- Engine/Pool:
- some new Pool utility classes, updated docs
- "use_threadlocal" on Pool defaults to False (same as create_engine)
- fixed direct execution of Compiled objects
- create_engine() reworked to be strict about incoming **kwargs. all keyword
arguments must be consumed by one of the dialect, connection pool, and engine
constructors, else a TypeError is thrown which describes the full set of
invalid kwargs in relation to the selected dialect/pool/engine configuration.
- Databases/Types:
- MySQL catches exception on "describe" and reports as NoSuchTableError
- further fixes to sqlite booleans, weren't working as defaults
- fix to postgres sequence quoting when using schemas
- ORM:
- the "delete" cascade will load in all child objects, if they were not
loaded already. this can be turned off (i.e. the old behavior) by setting
passive_deletes=True on a relation().
- adjustments to reworked eager query generation to not fail on circular
eager-loaded relationships (like backrefs)
- fixed bug where eagerload() (nor lazyload()) option didn't properly
instruct the Query whether or not to use "nesting" when producing a
LIMIT query.
- fixed bug in circular dependency sorting at flush time; if object A
contained a cyclical many-to-one relationship to object B, and object B
was just attached to object A, *but* object B itself wasnt changed,
the many-to-one synchronize of B's primary key attribute to A's foreign key
attribute wouldnt occur. [ticket:360]
- implemented from_obj argument for query.count, improves count function
on selectresults [ticket:325]
- added an assertion within the "cascade" step of ORM relationships to check
that the class of object attached to a parent object is appropriate
(i.e. if A.items stores B objects, raise an error if a C is appended to A.items)
- new extension sqlalchemy.ext.associationproxy, provides transparent
"association object" mappings. new example
examples/association/proxied_association.py illustrates.
- improvement to single table inheritance to load full hierarchies beneath
the target class
- fix to subtle condition in topological sort where a node could appear twice,
for [ticket:362]
- additional rework to topological sort, refactoring, for [ticket:365]
- "delete-orphan" for a certain type can be set on more than one parent class;
the instance is an "orphan" only if its not attached to *any* of those parents
0.3.0
- General:
- logging is now implemented via standard python "logging" module.
"echo" keyword parameters are still functional but set/unset
log levels for their respective classes/instances. all logging
can be controlled directly through the Python API by setting
INFO and DEBUG levels for loggers in the "sqlalchemy" namespace.
class-level logging is under "sqlalchemy.<module>.<classname>",
instance-level logging under "sqlalchemy.<module>.<classname>.0x..<00-FF>".
Test suite includes "--log-info" and "--log-debug" arguments
which work independently of --verbose/--quiet. Logging added
to orm to allow tracking of mapper configurations, row iteration.
- the documentation-generation system has been overhauled to be
much simpler in design and more integrated with Markdown
- Specific Databases:
- SQLite:
- sqlite boolean datatype converts False/True to 0/1 by default
- fixes to Date/Time (SLDate/SLTime) types; works as good as postgres
now [ticket:335]
- MS-SQL:
- fixes bug 261 (table reflection broken for MS-SQL case-sensitive
databases)
- can now specify port for pymssql
- introduces new "auto_identity_insert" option for auto-switching
between "SET IDENTITY_INSERT" mode when values specified for IDENTITY columns
- now supports multi-column foreign keys
- fix to reflecting date/datetime columns
- NCHAR and NVARCHAR type support added
- Oracle:
- Oracle has experimental support for cx_Oracle.TIMESTAMP, which requires
a setinputsizes() call on the cursor that is now enabled via the
'auto_setinputsizes' flag to the oracle dialect.
- Firebird:
- aliases do not use "AS"
- correctly raises NoSuchTableError when reflecting non-existent table
- Schema:
- a fair amount of cleanup to the schema package, removal of ambiguous
methods, methods that are no longer needed. slightly more constrained
useage, greater emphasis on explicitness
- the "primary_key" attribute of Table and other selectables becomes
a setlike ColumnCollection object; is ordered but not numerically
indexed. a comparison clause between two pks that are derived from the
same underlying tables (i.e. such as two Alias objects) can be generated
via table1.primary_key==table2.primary_key
- ForeignKey(Constraint) supports "use_alter=True", to create/drop a foreign key
via ALTER. this allows circular foreign key relationships to be set up.
- append_item() methods removed from Table and Column; preferably
construct Table/Column/related objects inline, but if needed use
append_column(), append_foreign_key(), append_constraint(), etc.
- table.create() no longer returns the Table object, instead has no
return value. the usual case is that tables are created via metadata,
which is preferable since it will handle table dependencies.
- added UniqueConstraint (goes at Table level), CheckConstraint
(goes at Table or Column level).
- index=False/unique=True on Column now creates a UniqueConstraint,
index=True/unique=False creates a plain Index,
index=True/unique=True on Column creates a unique Index. 'index'
and 'unique' keyword arguments to column are now boolean only; for
explcit names and groupings of indexes or unique constraints, use the
UniqueConstraint/Index constructs explicitly.
- added autoincrement=True to Column; will disable schema generation
of SERIAL/AUTO_INCREMENT/identity seq for postgres/mysql/mssql if
explicitly set to False
- TypeEngine objects now have methods to deal with copying and comparing
values of their specific type. Currently used by the ORM, see below.
- fixed condition that occurred during reflection when a primary key
column was explciitly overridden, where the PrimaryKeyConstraint would
get both the reflected and the programmatic column doubled up
- the "foreign_key" attribute on Column and ColumnElement in general
is deprecated, in favor of the "foreign_keys" list/set-based attribute,
which takes into account multiple foreign keys on one column.
"foreign_key" will return the first element in the "foreign_keys" list/set
or None if the list is empty.
- Connections/Pooling/Execution:
- connection pool tracks open cursors and automatically closes them
if connection is returned to pool with cursors still opened. Can be
affected by options which cause it to raise an error instead, or to
do nothing. fixes issues with MySQL, others
- fixed bug where Connection wouldnt lose its Transaction
after commit/rollback
- added scalar() method to ComposedSQLEngine, ResultProxy
- ResultProxy will close() the underlying cursor when the ResultProxy
itself is closed. this will auto-close cursors for ResultProxy objects
that have had all their rows fetched (or had scalar() called).
- ResultProxy.fetchall() internally uses DBAPI fetchall() for better efficiency,
added to mapper iteration as well (courtesy Michael Twomey)
- SQL Construction:
- changed "for_update" parameter to accept False/True/"nowait"
and "read", the latter two of which are interpreted only by
Oracle and Mysql [ticket:292]
- added extract() function to sql dialect
(SELECT extract(field FROM expr))
- BooleanExpression includes new "negate" argument to specify
the appropriate negation operator if one is available.
- calling a negation on an "IN" or "IS" clause will result in
"NOT IN", "IS NOT" (as opposed to NOT (x IN y)).
- Function objects know what to do in a FROM clause now. their
behavior should be the same, except now you can also do things like
select(['*'], from_obj=[func.my_function()]) to get multiple
columns from the result, or even use sql.column() constructs to name the
return columns [ticket:172]
- ORM:
- attribute tracking modified to be more intelligent about detecting
changes, particularly with mutable types. TypeEngine objects now
take a greater role in defining how to compare two scalar instances,
including the addition of a MutableType mixin which is implemented by
PickleType. unit-of-work now tracks the "dirty" list as an expression
of all persistent objects where the attribute manager detects changes.
The basic issue thats fixed is detecting changes on PickleType
objects, but also generalizes type handling and "modified" object
checking to be more complete and extensible.
- a wide refactoring to "attribute loader" and "options" architectures.
ColumnProperty and PropertyLoader define their loading behaivor via switchable
"strategies", and MapperOptions no longer use mapper/property copying
in order to function; they are instead propagated via QueryContext
and SelectionContext objects at query/instances time.
All of the internal copying of mappers and properties that was used to handle
inheritance as well as options() has been removed; the structure
of mappers and properties is much simpler than before and is clearly laid out
in the new 'interfaces' module.
- related to the mapper/property overhaul, internal refactoring to
mapper instances() method to use a SelectionContext object to track
state during the operation.
SLIGHT API BREAKAGE: the append_result() and populate_instances()
methods on MapperExtension have a slightly different method signature
now as a result of the change; hoping that these methods are not
in widespread use as of yet.
- instances() method moved to Query now, backwards-compatible
version remains on Mapper.
- added contains_eager() MapperOption, used in conjunction with
instances() to specify properties that should be eagerly loaded
from the result set, using their plain column names by default, or translated
given an custom row-translation function.
- more rearrangements of unit-of-work commit scheme to better allow
dependencies within circular flushes to work properly...updated
task traversal/logging implementation
- polymorphic mappers (i.e. using inheritance) now produces INSERT
statements in order of tables across all inherited classes
[ticket:321]
- added an automatic "row switch" feature to mapping, which will
detect a pending instance/deleted instance pair with the same
identity key and convert the INSERT/DELETE to a single UPDATE
- "association" mappings simplified to take advantage of
automatic "row switch" feature
- "custom list classes" is now implemented via the "collection_class"
keyword argument to relation(). the old way still works but is
deprecated [ticket:212]
- added "viewonly" flag to relation(), allows construction of
relations that have no effect on the flush() process.
- added "lockmode" argument to base Query select/get functions,
including "with_lockmode" function to get a Query copy that has
a default locking mode. Will translate "read"/"update"
arguments into a for_update argument on the select side.
[ticket:292]
- implemented "version check" logic in Query/Mapper, used
when version_id_col is in effect and query.with_lockmode()
is used to get() an instance thats already loaded
- post_update behavior improved; does a better job at not
updating too many rows, updates only required columns
[ticket:208]
- adjustments to eager loading so that its "eager chain" is
kept separate from the normal mapper setup, thereby
preventing conflicts with lazy loader operation, fixes
[ticket:308]
- fix to deferred group loading
- session.flush() wont close a connection it opened [ticket:346]
- added "batch=True" flag to mapper; if False, save_obj
will fully save one object at a time including calls
to before_XXXX and after_XXXX
- added "column_prefix=None" argument to mapper; prepends the
given string (typically '_') to column-based attributes automatically
set up from the mapper's Table
- specifying joins in the from_obj argument of query.select() will
replace the main table of the query, if the table is somewhere within
the given from_obj. this makes it possible to produce custom joins and
outerjoins in queries without the main table getting added twice.
[ticket:315]
- eagerloading is adjusted to more thoughtfully attach its LEFT OUTER JOINs
to the given query, looking for custom "FROM" clauses that may have
already been set up.
- added join_to and outerjoin_to transformative methods to SelectResults,
to build up join/outerjoin conditions based on property names. also
added select_from to explicitly set from_obj parameter.
- removed "is_primary" flag from mapper.
0.2.8
- cleanup on connection methods + documentation. custom DBAPI
arguments specified in query string, 'connect_args' argument
to 'create_engine', or custom creation function via 'creator'
function to 'create_engine'.
- added "recycle" argument to Pool, is "pool_recycle" on create_engine,
defaults to 3600 seconds; connections after this age will be closed and
replaced with a new one, to handle db's that automatically close
stale connections [ticket:274]
- changed "invalidate" semantics with pooled connection; will
instruct the underlying connection record to reconnect the next
time its called. "invalidate" will also automatically be called
if any error is thrown in the underlying call to connection.cursor().
this will hopefully allow the connection pool to reconnect to a
database that had been stopped and started without restarting
the connecting application [ticket:121]
- eesh ! the tutorial doctest was broken for quite some time.
- add_property() method on mapper does a "compile all mappers"
step in case the given property references a non-compiled mapper
(as it did in the case of the tutorial !)
- [ticket:277] check for pg sequence already existing before create
- if a contextual session is established via MapperExtension.get_session
(as it is using the sessioncontext plugin, etc), a lazy load operation
will use that session by default if the parent object is not
persistent with a session already.
- lazy loads will not fire off for an object that does not have a
database identity (why?
see http://www.sqlalchemy.org/trac/wiki/WhyDontForeignKeysLoadData)
- unit-of-work does a better check for "orphaned" objects that are
part of a "delete-orphan" cascade, for certain conditions where the
parent isnt available to cascade from.
- mappers can tell if one of their objects is an "orphan" based
on interactions with the attribute package. this check is based
on a status flag maintained for each relationship
when objects are attached and detached from each other.
- it is now invalid to declare a self-referential relationship with
"delete-orphan" (as the abovementioned check would make them impossible
to save)
- improved the check for objects being part of a session when the
unit of work seeks to flush() them as part of a relationship..
- [ticket:280] statement execution supports using the same BindParam
object more than once in an expression; simplified handling of positional
parameters. nice job by Bill Noon figuring out the basic idea.
- postgres reflection moved to use pg_schema tables, can be overridden
with use_information_schema=True argument to create_engine
[ticket:60], [ticket:71]
- added case_sensitive argument to MetaData, Table, Column, determines
itself automatically based on if a parent schemaitem has a non-None
setting for the flag, or if not, then whether the identifier name is all lower
case or not. when set to True, quoting is applied to identifiers with mixed or
uppercase identifiers. quoting is also applied automatically in all cases to
identifiers that are known to be reserved words or contain other non-standard
characters. various database dialects can override all of this behavior, but
currently they are all using the default behavior. tested with postgres, mysql,
sqlite, oracle. needs more testing with firebird, ms-sql. part of the ongoing
work with [ticket:155]
- unit tests updated to run without any pysqlite installed; pool
test uses a mock DBAPI
- urls support escaped characters in passwords [ticket:281]
- added limit/offset to UNION queries (though not yet in oracle)
- added "timezone=True" flag to DateTime and Time types. postgres
so far will convert this to "TIME[STAMP] (WITH|WITHOUT) TIME ZONE",
so that control over timezone presence is more controllable (psycopg2
returns datetimes with tzinfo's if available, which can create confusion
against datetimes that dont).
- fix to using query.count() with distinct, **kwargs with SelectResults
count() [ticket:287]
- deregister Table from MetaData when autoload fails; [ticket:289]
- import of py2.5s sqlite3 [ticket:293]
- unicode fix for startswith()/endswith() [ticket:296]
0.2.7
- quoting facilities set up so that database-specific quoting can be
turned on for individual table, schema, and column identifiers when
used in all queries/creates/drops. Enabled via "quote=True" in
Table or Column, as well as "quote_schema=True" in Table. Thanks to
Aaron Spike for his excellent efforts.
- assignmapper was setting is_primary=True, causing all sorts of mayhem
by not raising an error when redundant mappers were set up, fixed
- added allow_null_pks option to Mapper, allows rows where some
primary key columns are null (i.e. when mapping to outer joins etc)
- modifcation to unitofwork to not maintain ordering within the
"new" list or within the UOWTask "objects" list; instead, new objects
are tagged with an ordering identifier as they are registered as new
with the session, and the INSERT statements are then sorted within the
mapper save_obj. the INSERT ordering has basically been pushed all
the way to the end of the flush cycle. that way the various sorts and
organizations occuring within UOWTask (particularly the circular task
sort) dont have to worry about maintaining order (which they werent anyway)
- fixed reflection of foreign keys to autoload the referenced table
if it was not loaded already
- [ticket:256] - pass URL query string arguments to connect() function
- [ticket:257] - oracle boolean type
- custom primary/secondary join conditions in a relation *will* be propagated
to backrefs by default. specifying a backref() will override this behavior.
- better check for ambiguous join conditions in sql.Join; propagates to a
better error message in PropertyLoader (i.e. relation()/backref()) for when
the join condition can't be reasonably determined.
- sqlite creates ForeignKeyConstraint objects properly upon table
reflection.
- adjustments to pool stemming from changes made for [ticket:224].
overflow counter should only be decremented if the connection actually
succeeded. added a test script to attempt testing this.
- fixed mysql reflection of default values to be PassiveDefault
- added reflected 'tinyint', 'mediumint' type to MS-SQL [ticket:263],
[ticket:264]
- SingletonThreadPool has a size and does a cleanup pass, so that
only a given number of thread-local connections stay around (needed
for sqlite applications that dispose of threads en masse)
- fixed small pickle bug(s) with lazy loaders [ticket:265] [ticket:267]
- fixed possible error in mysql reflection where certain versions
return an array instead of string for SHOW CREATE TABLE call
- fix to lazy loads when mapping to joins [changeset:1770]
- all create()/drop() calls have a keyword argument of "connectable".
"engine" is deprecated.
- fixed ms-sql connect() to work with adodbapi
- added "nowait" flag to Select()
- inheritance check uses issubclass() instead of direct __mro__ check
to make sure class A inherits from B, allowing mapper inheritance to more
flexibly correspond to class inheritance [ticket:271]
- SelectResults will use a subselect, when calling an aggregate (i.e.
max, min, etc.) on a SelectResults that has an ORDER BY clause
[ticket:252]
- fixes to types so that database-specific types more easily used;
fixes to mysql text types to work with this methodology
[ticket:269]
- some fixes to sqlite date type organization
- added MSTinyInteger to MS-SQL [ticket:263]
0.2.6
- big overhaul to schema to allow truly composite primary and foreign
key constraints, via new ForeignKeyConstraint and PrimaryKeyConstraint
objects.
Existing methods of primary/foreign key creation have not been changed
but use these new objects behind the scenes. table creation
and reflection is now more table oriented rather than column oriented.
[ticket:76]
- overhaul to MapperExtension calling scheme, wasnt working very well
previously
- tweaks to ActiveMapper, supports self-referential relationships
- slight rearrangement to objectstore (in activemapper/threadlocal)
so that the SessionContext is referenced by '.context' instead
of subclassed directly.
- activemapper will use threadlocal's objectstore if the mod is
activated when activemapper is imported
- small fix to URL regexp to allow filenames with '@' in them
- fixes to Session expunge/update/etc...needs more cleanup.
- select_table mappers *still* werent always compiling
- fixed up Boolean datatype
- added count()/count_by() to list of methods proxied by assignmapper;
this also adds them to activemapper
- connection exceptions wrapped in DBAPIError
- ActiveMapper now supports autoloading column definitions from the
database if you supply a __autoload__ = True attribute in your
mapping inner-class. Currently this does not support reflecting
any relationships.
- deferred column load could screw up the connection status in
a flush() under some circumstances, this was fixed
- expunge() was not working with cascade, fixed.
- potential endless loop in cascading operations fixed.
- added "synonym()" function, applied to properties to have a
propname the same as another, for the purposes of overriding props
and allowing the original propname to be accessible in select_by().
- fix to typing in clause construction which specifically helps
type issues with polymorphic_union (CAST/ColumnClause propagates
its type to proxy columns)
- mapper compilation work ongoing, someday it'll work....moved
around the initialization of MapperProperty objects to be after
all mappers are created to better handle circular compilations.
do_init() method is called on all properties now which are more
aware of their "inherited" status if so.
- eager loads explicitly disallowed on self-referential relationships, or
relationships to an inheriting mapper (which is also self-referential)
- reduced bind param size in query._get to appease the picky oracle
[ticket:244]
- added 'checkfirst' argument to table.create()/table.drop(), as
well as table.exists() [ticket:234]
- some other ongoing fixes to inheritance [ticket:245]
- attribute/backref/orphan/history-tracking tweaks as usual...
0.2.5
- fixed endless loop bug in select_by(), if the traversal hit
two mappers that referenced each other
- upgraded all unittests to insert './lib/' into sys.path,
working around new setuptools PYTHONPATH-killing behavior
- further fixes with attributes/dependencies/etc....
- improved error handling for when DynamicMetaData is not connected
- MS-SQL support largely working (tested with pymssql)
- ordering of UPDATE and DELETE statements within groups is now
in order of primary key values, for more deterministic ordering
- after_insert/delete/update mapper extensions now called per object,
not per-object-per-table
- further fixes/refactorings to mapper compilation
0.2.4
- try/except when the mapper sets init.__name__ on a mapped class,
supports python 2.3
- fixed bug where threadlocal engine would still autocommit
despite a transaction in progress
- lazy load and deferred load operations require the parent object
to be in a Session to do the operation; whereas before the operation
would just return a blank list or None, it now raises an exception.
- Session.update() is slightly more lenient if the session to which
the given object was formerly attached to was garbage collected;
otherwise still requires you explicitly remove the instance from
the previous Session.
- fixes to mapper compilation, checking for more error conditions
- small fix to eager loading combined with ordering/limit/offset
- utterly remarkable: added a single space between 'CREATE TABLE'
and '(<the rest of it>' since *thats how MySQL indicates a non-
reserved word tablename.....* [ticket:206]
- more fixes to inheritance, related to many-to-many relations
properly saving
- fixed bug when specifying explicit module to mysql dialect
- when QueuePool times out it raises a TimeoutError instead of
erroneously making another connection
- Queue.Queue usage in pool has been replaced with a locally
modified version (works in py2.3/2.4!) that uses a threading.RLock
for a mutex. this is to fix a reported case where a ConnectionFairy's
__del__() method got called within the Queue's get() method, which
then returns its connection to the Queue via the the put() method,
causing a reentrant hang unless threading.RLock is used.
- postgres will not place SERIAL keyword on a primary key column
if it has a foreign key constraint
- cursor() method on ConnectionFairy allows db-specific extension
arguments to be propagated [ticket:221]
- lazy load bind params properly propagate column type [ticket:225]
- new MySQL types: MSEnum, MSTinyText, MSMediumText, MSLongText, etc.
more support for MS-specific length/precision params in numeric types
patch courtesy Mike Bernson
- some fixes to connection pool invalidate() [ticket:224]
0.2.3
- overhaul to mapper compilation to be deferred. this allows mappers
to be constructed in any order, and their relationships to each
other are compiled when the mappers are first used.
- fixed a pretty big speed bottleneck in cascading behavior particularly
when backrefs were in use
- the attribute instrumentation module has been completely rewritten; its
now a large degree simpler and clearer, slightly faster. the "history"
of an attribute is no longer micromanaged with each change and is
instead part of a "CommittedState" object created when the
instance is first loaded. HistoryArraySet is gone, the behavior of
list attributes is now more open ended (i.e. theyre not sets anymore).
- py2.4 "set" construct used internally, falls back to sets.Set when
"set" not available/ordering is needed.
- fix to transaction control, so that repeated rollback() calls
dont fail (was failing pretty badly when flush() would raise
an exception in a larger try/except transaction block)
- "foreignkey" argument to relation() can also be a list. fixed
auto-foreignkey detection [ticket:151]
- fixed bug where tables with schema names werent getting indexed in
the MetaData object properly
- fixed bug where Column with redefined "key" property wasnt getting
type conversion happening in the ResultProxy [ticket:207]
- fixed 'port' attribute of URL to be an integer if present
- fixed old bug where if a many-to-many table mapped as "secondary"
had extra columns, delete operations didnt work
- bugfixes for mapping against UNION queries
- fixed incorrect exception class thrown when no DB driver present
- added NonExistentTable exception thrown when reflecting a table
that doesnt exist [ticket:138]
- small fix to ActiveMapper regarding one-to-one backrefs, other
refactorings
- overridden constructor in mapped classes gets __name__ and
__doc__ from the original class
- fixed small bug in selectresult.py regarding mapper extension
[ticket:200]
- small tweak to cascade_mappers, not very strongly supported
function at the moment
- some fixes to between(), column.between() to propagate typing
information better [ticket:202]
- if an object fails to be constructed, is not added to the
session [ticket:203]
- CAST function has been made into its own clause object with
its own compilation function in ansicompiler; allows MySQL
to silently ignore most CAST calls since MySQL
seems to only support the standard CAST syntax with Date types.
MySQL-compatible CAST support for strings, ints, etc. a TODO
0.2.2
- big improvements to polymorphic inheritance behavior, enabling it
to work with adjacency list table structures [ticket:190]
- major fixes and refactorings to inheritance relationships overall,
more unit tests
- fixed "echo_pool" flag on create_engine()
- fix to docs, removed incorrect info that close() is unsafe to use
with threadlocal strategy (its totally safe !)
- create_engine() can take URLs as string or unicode [ticket:188]
- firebird support partially completed;
thanks to James Ralston and Brad Clements for their efforts.
- Oracle url translation was broken, fixed, will feed host/port/sid
into cx_oracle makedsn() if 'database' field is present, else uses
straight TNS name from the 'host' field
- fix to using unicode criterion for query.get()/query.load()
- count() function on selectables now uses table primary key or
first column instead of "1" for criterion, also uses label "rowcount"
instead of "count".
- got rudimental "mapping to multiple tables" functionality cleaned up,
more correctly documented
- restored global_connect() function, attaches to a DynamicMetaData
instance called "default_metadata". leaving MetaData arg to Table
out will use the default metadata.
- fixes to session cascade behavior, entity_name propigation
- reorganized unittests into subdirectories
- more fixes to threadlocal connection nesting patterns
0.2.1
- "pool" argument to create_engine() properly propagates
- fixes to URL, raises exception if not parsed, does not pass blank
fields along to the DB connect string (a string such as
user:host@/db was breaking on postgres)
- small fixes to Mapper when it inserts and tries to get
new primary key values back
- rewrote half of TLEngine, the ComposedSQLEngine used with
'strategy="threadlocal"'. it now properly implements engine.begin()/
engine.commit(), which nest fully with connection.begin()/trans.commit().
added about six unittests.
- major "duh" in pool.Pool, forgot to put back the WeakValueDictionary.
unittest which was supposed to check for this was also silently missing
it. fixed unittest to ensure that ConnectionFairy properly falls out
of scope.
- placeholder dispose() method added to SingletonThreadPool, doesnt
do anything yet
- rollback() is automatically called when an exception is raised,
but only if theres no transaction in process (i.e. works more like
autocommit).
- fixed exception raise in sqlite if no sqlite module present
- added extra example detail for association object doc
- Connection adds checks for already being closed
0.2.0
- overhaul to Engine system so that what was formerly the SQLEngine
is now a ComposedSQLEngine which consists of a variety of components,
including a Dialect, ConnectionProvider, etc. This impacted all the
db modules as well as Session and Mapper.
- create_engine now takes only RFC-1738-style strings:
driver://user:password@host:port/database
- total rewrite of connection-scoping methodology, Connection objects
can now execute clause elements directly, added explicit "close" as
well as support throughout Engine/ORM to handle closing properly,
no longer relying upon __del__ internally to return connections
to the pool [ticket:152].
- overhaul to Session interface and scoping. uses hibernate-style
methods, including query(class), save(), save_or_update(), etc.
no threadlocal scope is installed by default. Provides a binding
interface to specific Engines and/or Connections so that underlying
Schema objects do not need to be bound to an Engine. Added a basic
SessionTransaction object that can simplistically aggregate transactions
across multiple engines.
- overhaul to mapper's dependency and "cascade" behavior; dependency logic
factored out of properties.py into a separate module "dependency.py".
"cascade" behavior is now explicitly controllable, proper implementation
of "delete", "delete-orphan", etc. dependency system can now determine at
flush time if a child object has a parent or not so that it makes better
decisions on how that child should be updated in the DB with regards to deletes.
- overhaul to Schema to build upon MetaData object instead of an Engine.
Entire SQL/Schema system can be used with no Engines whatsoever, executed
solely by an explicit Connection object. the "bound" methodlogy exists via the
BoundMetaData for schema objects. ProxyEngine is generally not needed
anymore and is replaced by DynamicMetaData.
- true polymorphic behavior implemented, fixes [ticket:167]
- "oid" system has been totally moved into compile-time behavior;
if they are used in an order_by where they are not available, the order_by
doesnt get compiled, fixes [ticket:147]
- overhaul to packaging; "mapping" is now "orm", "objectstore" is now
"session", the old "objectstore" namespace gets loaded in via the
"threadlocal" mod if used
- mods now called in via "import <modname>". extensions favored over
mods as mods are globally-monkeypatching
- fix to add_property so that it propagates properties to inheriting
mappers [ticket:154]
- backrefs create themselves against primary mapper of its originating
property, priamry/secondary join arguments can be specified to override.
helps their usage with polymorphic mappers
- "table exists" function has been implemented [ticket:31]
- "create_all/drop_all" added to MetaData object [ticket:98]
- improvements and fixes to topological sort algorithm, as well as more
unit tests
- tutorial page added to docs which also can be run with a custom doctest
runner to ensure its properly working. docs generally overhauled to
deal with new code patterns
- many more fixes, refactorings.
- migration guide is available on the Wiki at
http://www.sqlalchemy.org/trac/wiki/02Migration
0.1.7
- some fixes to topological sort algorithm
- added DISTINCT ON support to Postgres (just supply distinct=[col1,col2..])
- added __mod__ (% operator) to sql expressions
- "order_by" mapper property inherited from inheriting mapper
- fix to column type used when mapper UPDATES/DELETEs
- with convert_unicode=True, reflection was failing, has been fixed
- types types types! still werent working....have to use TypeDecorator again :(
- mysql binary type converts array output to buffer, fixes PickleType
- fixed the attributes.py memory leak once and for all
- unittests are qualified based on the databases that support each one
- fixed bug where column defaults would clobber VALUES clause of insert objects
- fixed bug where table def w/ schema name would force engine connection
- fix for parenthesis to work correctly with subqueries in INSERT/UPDATE
- HistoryArraySet gets extend() method
- fixed lazyload support for other comparison operators besides =
- lazyload fix where two comparisons in the join condition point to the
samem column
- added "construct_new" flag to mapper, will use __new__ to create instances
instead of __init__ (standard in 0.2)
- added selectresults.py to SVN, missed it last time
- tweak to allow a many-to-many relationship from a table to itself via
an association table
- small fix to "translate_row" function used by polymorphic example
- create_engine uses cgi.parse_qsl to read query string (out the window in 0.2)
- tweaks to CAST operator
- fixed function names LOCAL_TIME/LOCAL_TIMESTAMP -> LOCALTIME/LOCALTIMESTAMP
- fixed order of ORDER BY/HAVING in compile
0.1.6
- support for MS-SQL added courtesy Rick Morrison, Runar Petursson
- the latest SQLSoup from J. Ellis
- ActiveMapper has preliminary support for inheritance (Jeff Watkins)
- added a "mods" system which allows pluggable modules that modify/augment
core functionality, using the function "install_mods(*modnames)".
- added the first "mod", SelectResults, which modifies mapper selects to
return generators that turn ranges into LIMIT/OFFSET queries
(Jonas Borgstr?
- factored out querying capabilities of Mapper into a separate Query object
which is Session-centric. this improves the performance of mapper.using(session)
and makes other things possible.
- objectstore/Session refactored, the official way to save objects is now
via the flush() method. The begin/commit functionality of Session is factored
into LegacySession which is still established as the default behavior, until
the 0.2 series.
- types system is bound to an engine at query compile time, not schema
construction time. this simplifies the types system as well as the ProxyEngine.
- added 'version_id' keyword argument to mapper. this keyword should reference a
Column object with type Integer, preferably non-nullable, which will be used on
the mapped table to track version numbers. this number is incremented on each
save operation and is specifed in the UPDATE/DELETE conditions so that it
factors into the returned row count, which results in a ConcurrencyError if the
value received is not the expected count.
- added 'entity_name' keyword argument to mapper. a mapper is now associated
with a class via the class object as well as an optional entity_name parameter,
which is a string defaulting to None. any number of primary mappers can be
created for a class, qualified by the entity name. instances of those classes
will issue all of their load and save operations through their
entity_name-qualified mapper, and maintain separate a identity in the identity
map for an otherwise equilvalent object.
- overhaul to the attributes system. code has been clarified, and also fixed to
support proper polymorphic behavior on object attributes.
- added "for_update" flag to Select objects
- some fixes for backrefs
- fix for postgres1 DateTime type
- documentation pages mostly switched over to Markdown syntax
0.1.5
- added SQLSession concept to SQLEngine. this object keeps track of retrieving a
connection from the connection pool as well as an in-progress transaction.
methods push_session() and pop_session() added to SQLEngine which push/pop a new
SQLSession onto the engine, allowing operation upon a second connection "nested"
within the previous one, allowing nested transactions. Other tricks are sure to
come later regarding SQLSession.
- added nest_on argument to objectstore.Session. This is a single SQLEngine or
list of engines for which push_session()/pop_session() will be called each time
this Session becomes the active session (via objectstore.push_session() or
equivalent). This allows a unit of work Session to take advantage of the nested
transaction feature without explicitly calling push_session/pop_session on the
engine.
- factored apart objectstore/unitofwork to separate "Session scoping" from
"uow commit heavy lifting"
- added populate_instance() method to MapperExtension. allows an extension to
modify the population of object attributes. this method can call the
populate_instance() method on another mapper to proxy the attribute population
from one mapper to another; some row translation logic is also built in to help
with this.
- fixed Oracle8-compatibility "use_ansi" flag which converts JOINs to
comparisons with the = and (+) operators, passes basic unittests
- tweaks to Oracle LIMIT/OFFSET support
- Oracle reflection uses ALL_** views instead of USER_** to get larger
list of stuff to reflect from
- fixes to Oracle foreign key reflection [ticket:105]
- objectstore.commit(obj1, obj2,...) adds an extra step to seek out private
relations on properties and delete child objects, even though its not a global
commit
- lots and lots of fixes to mappers which use inheritance, strengthened the
concept of relations on a mapper being made towards the "local" table for that
mapper, not the tables it inherits. allows more complex compositional patterns
to work with lazy/eager loading.
- added support for mappers to inherit from others based on the same table,
just specify the same table as that of both parent/child mapper.
- some minor speed improvements to the attributes system with regards to
instantiating and populating new objects.
- fixed MySQL binary unit test
- INSERTs can receive clause elements as VALUES arguments, not just literal
values
- support for calling multi-tokened functions, i.e. schema.mypkg.func()
- added J. Ellis' SQLSoup module to extensions package
- added "polymorphic" examples illustrating methods to load multiple object types
from one mapper, the second of which uses the new populate_instance() method.
small improvements to mapper, UNION construct to help the examples along
- improvements/fixes to session.refresh()/session.expire() (which may have
been called "invalidate" earlier..)
- added session.expunge() which totally removes an object from the current
session
- added *args, **kwargs pass-thru to engine.transaction(func) allowing easier
creation of transactionalizing decorator functions
- added iterator interface to ResultProxy: "for row in result:..."
- added assertion to tx = session.begin(); tx.rollback(); tx.begin(), i.e. cant
use it after a rollback()
- added date conversion on bind parameter fix to SQLite enabling dates to
work with pysqlite1
- improvements to subqueries to more intelligently construct their FROM
clauses [ticket:116]
- added PickleType to types.
- fixed two bugs with column labels with regards to bind parameters: bind param
keynames they are now generated from a column "label" in all relevant cases to
take advantage of excess-name-length rules, and checks for a peculiar collision
against a column named the same as "tablename_colname" added
- major overhaul to unit of work documentation, other documentation sections.
- fixed attributes bug where if an object is committed, its lazy-loaded list got
blown away if it hadnt been loaded
- added unique_connection() method to engine, connection pool to return a
connection that is not part of the thread-local context or any current
transaction
- added invalidate() function to pooled connection. will remove the connection
from the pool. still need work for engines to auto-reconnect to a stale DB
though.
- added distinct() function to column elements so you can do
func.count(mycol.distinct())
- added "always_refresh" flag to Mapper, creates a mapper that will always
refresh the attributes of objects it gets/selects from the DB, overwriting any
changes made.
0.1.4
- create_engine() now uses genericized parameters; host/hostname,
db/dbname/database, password/passwd, etc. for all engine connections. makes
engine URIs much more "universal"
- added support for SELECT statements embedded into a column clause, using the
flag "scalar=True"
- another overhaul to EagerLoading when used in conjunction with mappers that
inherit; improvements to eager loads figuring out their aliased queries
correctly, also relations set up against a mapper with inherited mappers will
create joins against the table that is specific to the mapper itself (i.e. and
not any tables that are inherited/are further down the inheritance chain),
this can be overridden by using custom primary/secondary joins.
- added J.Ellis patch to mapper.py so that selectone() throws an exception
if query returns more than one object row, selectfirst() to not throw the
exception. also adds selectfirst_by (synonymous with get_by) and selectone_by
- added onupdate parameter to Column, will exec SQL/python upon an update
statement.Also adds "for_update=True" to all DefaultGenerator subclasses
- added support for Oracle table reflection contributed by Andrija Zaric;
still some bugs to work out regarding composite primary keys/dictionary selection
- checked in an initial Firebird module, awaiting testing.
- added sql.ClauseParameters dictionary object as the result for
compiled.get_params(), does late-typeprocessing of bind parameters so
that the original values are easier to access
- more docs for indexes, column defaults, connection pooling, engine construction
- overhaul to the construction of the types system. uses a simpler inheritance
pattern so that any of the generic types can be easily subclassed, with no need
for TypeDecorator.
- added "convert_unicode=False" parameter to SQLEngine, will cause all String
types to perform unicode encoding/decoding (makes Strings act like Unicodes)
- added 'encoding="utf8"' parameter to engine. the given encoding will be
used for all encode/decode calls within Unicode types as well as Strings
when convert_unicode=True.
- improved support for mapping against UNIONs, added polymorph.py example
to illustrate multi-class mapping against a UNION
- fix to SQLite LIMIT/OFFSET syntax
- fix to Oracle LIMIT syntax
- added backref() function, allows backreferences to have keyword arguments
that will be passed to the backref.
- Sequences and ColumnDefault objects can do execute()/scalar() standalone
- SQL functions (i.e. func.foo()) can do execute()/scalar() standalone
- fix to SQL functions so that the ANSI-standard functions, i.e. current_timestamp
etc., do not specify parenthesis. all other functions do.
- added settattr_clean and append_clean to SmartProperty, which set
attributes without triggering a "dirty" event or any history. used as:
myclass.prop1.setattr_clean(myobject, 'hi')
- improved support to column defaults when used by mappers; mappers will pull
pre-executed defaults from statement's executed bind parameters
(pre-conversion) to populate them into a saved object's attributes; if any
PassiveDefaults have fired off, will instead post-fetch the row from the DB to
populate the object.
- added 'get_session().invalidate(*obj)' method to objectstore, instances will
refresh() themselves upon the next attribute access.
- improvements to SQL func calls including an "engine" keyword argument so
they can be execute()d or scalar()ed standalone, also added func accessor to
SQLEngine
- fix to MySQL4 custom table engines, i.e. TYPE instead of ENGINE
- slightly enhanced logging, includes timestamps and a somewhat configurable
formatting system, in lieu of a full-blown logging system
- improvements to the ActiveMapper class from the TG gang, including
many-to-many relationships
- added Double and TinyInt support to mysql
0.1.3
- completed "post_update" feature, will add a second update statement before
inserts and after deletes in order to reconcile a relationship without any
dependencies being created; used when persisting two rows that are dependent
on each other
- completed mapper.using(session) function, localized per-object Session
functionality; objects can be declared and manipulated as local to any
user-defined Session
- fix to Oracle "row_number over" clause with multiple tables
- mapper.get() was not selecting multiple-keyed objects if the mapper's table was a join,
such as in an inheritance relationship, this is fixed.
- overhaul to sql/schema packages so that the sql package can run all on its own,
producing selects, inserts, etc. without any engine dependencies. builds upon
new TableClause/ColumnClause lexical objects. Schema's Table/Column objects
are the "physical" subclasses of them. simplifies schema/sql relationship,
extensions (like proxyengine), and speeds overall performance by a large margin.
removes the entire getattr() behavior that plagued 0.1.1.
- refactoring of how the mapper "synchronizes" data between two objects into a
separate module, works better with properties attached to a mapper that has an
additional inheritance relationship to one of the related tables, also the same
methodology used to synchronize parent/child objects now used by mapper to
synchronize between inherited and inheriting mappers.
- made objectstore "check for out-of-identitymap" more aggressive, will perform the
check when object attributes are modified or the object is deleted
- Index object fully implemented, can be constructed standalone, or via
"index" and "unique" arguments on Columns.
- added "convert_unicode" flag to SQLEngine, will treat all String/CHAR types
as Unicode types, with raw-byte/utf-8 translation on the bind parameter and
result set side.
- postgres maintains a list of ANSI functions that must have no parenthesis so
function calls with no arguments work consistently
- tables can be created with no engine specified. this will default their engine
to a module-scoped "default engine" which is a ProxyEngine. this engine can
be connected via the function "global_connect".
- added "refresh(*obj)" method to objectstore / Session to reload the attributes of
any set of objects from the database unconditionally
0.1.2
- fixed a recursive call in schema that was somehow running 994 times then returning
normally. broke nothing, slowed down everything. thanks to jpellerin for finding this.
0.1.1
- small fix to Function class so that expressions with a func.foo() use the type of the
Function object (i.e. the left side) as the type of the boolean expression, not the
other side which is more of a moving target (changeset 1020).
- creating self-referring mappers with backrefs slightly easier (but still not that easy -
changeset 1019)
- fixes to one-to-one mappings (changeset 1015)
- psycopg1 date/time issue with None fixed (changeset 1005)
- two issues related to postgres, which doesnt want to give you the "lastrowid"
since oids are deprecated:
* postgres database-side defaults that are on primary key cols *do* execute
explicitly beforehand, even though thats not the idea of a PassiveDefault. this is
because sequences on columns get reflected as PassiveDefaults, but need to be explicitly
executed on a primary key col so we know what we just inserted.
* if you did add a row that has a bunch of database-side defaults on it,
and the PassiveDefault thing was working the old way, i.e. they just execute on
the DB side, the "cant get the row back without an OID" exception that occurred
also will not happen unless someone (usually the ORM) explicitly asks for it.
- fixed a glitch with engine.execute_compiled where it was making a second
ResultProxy that just got thrown away.
- began to implement newer logic in object properities. you can now say
myclass.attr.property, which will give you the PropertyLoader corresponding to that
attribute, i.e. myclass.mapper.props['attr']
- eager loading has been internally overhauled to use aliases at all times. more
complicated chains of eager loads can now be created without any need for explicit
"use aliases"-type instructions. EagerLoader code is also much simpler now.
- a new somewhat experimental flag "use_update" added to relations, indicates that
this relationship should be handled by a second UPDATE statement, either after a
primary INSERT or before a primary DELETE. handles circular row dependencies.
- added exceptions module, all raised exceptions (except for some
KeyError/AttributeError exceptions) descend from these classes.
- fix to date types with MySQL, returned timedelta converted to datetime.time
- two-phase objectstore.commit operations (i.e. begin/commit) now return a
transactional object (SessionTrans), to more clearly indicate transaction boundaries.
- Index object with create/drop support added to schema
- fix to postgres, where it will explicitly pre-execute a PassiveDefault on a table
if it is a primary key column, pursuant to the ongoing "we cant get inserted rows
back from postgres" issue
- change to information_schema query that gets back postgres table defs, now
uses explicit JOIN keyword, since one user had faster performance with 8.1
- fix to engine.process_defaults so it works correctly with a table that has
different column name/column keys (changset 982)
- a column can only be attached to one table - this is now asserted
- postgres time types descend from Time type
- fix to alltests so that it runs types test (now named testtypes)
- fix to Join object so that it correctly exports its foreign keys (cs 973)
- creating relationships against mappers that use inheritance fixed (cs 973)
0.1.0
initial release
Reference in New Issue View Git Blame Copy Permalink