Sqlalchemy

:released: November 12, 2022

.. change::
:tags: bug, sql
:tickets: 8790
:versions: 2.0.0b4

Fixed critical memory issue identified in cache key generation, where for
very large and complex ORM statements that make use of lots of ORM aliases
with subqueries, cache key generation could produce excessively large keys
that were orders of magnitude bigger than the statement itself. Much thanks
to Rollo Konig Brock for their very patient, long term help in finally
identifying this issue.

.. change::
:tags: bug, postgresql, mssql
:tickets: 8770
:versions: 2.0.0b4

For the PostgreSQL and SQL Server dialects only, adjusted the compiler so
that when rendering column expressions in the RETURNING clause, the "non
anon" label that's used in SELECT statements is suggested for SQL
expression elements that generate a label; the primary example is a SQL
function that may be emitting as part of the column's type, where the label
name should match the column's name by default. This restores a not-well
defined behavior that had changed in version 1.4.21 due to :ticket:`6718`,
:ticket:`6710`. The Oracle dialect has a different RETURNING implementation
and was not affected by this issue. Version 2.0 features an across the
board change for its widely expanded support of RETURNING on other
backends.


.. change::
:tags: bug, oracle

Fixed issue in the Oracle dialect where an INSERT statement that used
``insert(some_table).values(...).returning(some_table)`` against a full
:class:`.Table` object at once would fail to execute, raising an exception.

.. change::
:tags: bug, tests
:tickets: 8793
:versions: 2.0.0b4

Fixed issue where the ``--disable-asyncio`` parameter to the test suite
would fail to not actually run greenlet tests and would also not prevent
the suite from using a "wrapping" greenlet for the whole suite. This
parameter now ensures that no greenlet or asyncio use will occur within the
entire run when set.

.. change::
:tags: bug, tests

Adjusted the test suite which tests the Mypy plugin to accommodate for
changes in Mypy 0.990 regarding how it handles message output, which affect
how sys.path is interpreted when determining if notes and errors should be
printed for particular files. The change broke the test suite as the files
within the test directory itself no longer produced messaging when run
under the mypy API.

.. changelog::

:released: November 4, 2022

.. change::
:tags: bug, orm
:tickets: 8738
:versions: 2.0.0b3

Fixed issue in joined eager loading where an assertion fail would occur
with a particular combination of outer/inner joined eager loads, when
eager loading across three mappers where the middle mapper was
an inherited subclass mapper.


.. change::
:tags: bug, oracle
:tickets: 8708
:versions: 2.0.0b3

Fixed issue where bound parameter names, including those automatically
derived from similarly-named database columns, which contained characters
that normally require quoting with Oracle would not be escaped when using
"expanding parameters" with the Oracle dialect, causing execution errors.
The usual "quoting" for bound parameters used by the Oracle dialect is not
used with the "expanding parameters" architecture, so escaping for a large
range of characters is used instead, now using a list of characters/escapes
that are specific to Oracle.



.. change::
:tags: bug, orm
:tickets: 8721
:versions: 2.0.0b3

Fixed bug involving :class:`.Select` constructs, where combinations of
:meth:`.Select.select_from` with :meth:`.Select.join`, as well as when
using :meth:`.Select.join_from`, would cause the
:func:`_orm.with_loader_criteria` feature as well as the IN criteria needed
for single-table inheritance queries to not render, in cases where the
columns clause of the query did not explicitly include the left-hand side
entity of the JOIN. The correct entity is now transferred to the
:class:`.Join` object that's generated internally, so that the criteria
against the left side entity is correctly added.


.. change::
:tags: bug, mssql
:tickets: 8714
:versions: 2.0.0b3

Fixed issue with :meth:`.Inspector.has_table`, which when used against a
temporary table with the SQL Server dialect would fail on some Azure
variants, due to an unnecessary information schema query that is not
supported on those server versions. Pull request courtesy Mike Barry.

.. change::
:tags: bug, orm
:tickets: 8711
:versions: 2.0.0b3

An informative exception is now raised when the
:func:`_orm.with_loader_criteria` option is used as a loader option added
to a specific "loader path", such as when using it within
:meth:`.Load.options`. This use is not supported as
:func:`_orm.with_loader_criteria` is only intended to be used as a top
level loader option. Previously, an internal error would be generated.

.. change::
:tags: bug, oracle
:tickets: 8744
:versions: 2.0.0b3

Fixed issue where the ``nls_session_parameters`` view queried on first
connect in order to get the default decimal point character may not be
available depending on Oracle connection modes, and would therefore raise
an error. The approach to detecting decimal char has been simplified to
test a decimal value directly, instead of reading system views, which
works on any backend / driver.


.. change::
:tags: bug, orm
:tickets: 8753
:versions: 2.0.0b3

Improved "dictionary mode" for :meth:`_orm.Session.get` so that synonym
names which refer to primary key attribute names may be indicated in the
named dictionary.

.. change::
:tags: bug, engine, regression
:tickets: 8717
:versions: 2.0.0b3

Fixed issue where the :meth:`.PoolEvents.reset` event hook would not be be
called in all cases when a :class:`_engine.Connection` were closed and was
in the process of returning its DBAPI connection to the connection pool.

The scenario was when the :class:`_engine.Connection` had already emitted
``.rollback()`` on its DBAPI connection within the process of returning
the connection to the pool, where it would then instruct the connection
pool to forego doing its own "reset" to save on the additional method
call. However, this prevented custom pool reset schemes from being
used within this hook, as such hooks by definition are doing more than
just calling ``.rollback()``, and need to be invoked under all
circumstances. This was a regression that appeared in version 1.4.

For version 1.4, the :meth:`.PoolEvents.checkin` remains viable as an
alternate event hook to use for custom "reset" implementations. Version 2.0
will feature an improved version of :meth:`.PoolEvents.reset` which is
called for additional scenarios such as termination of asyncio connections,
and is also passed contextual information about the reset, to allow for
"custom connection reset" schemes which can respond to different reset
scenarios in different ways.

.. change::
:tags: bug, orm
:tickets: 8704
:versions: 2.0.0b3

Fixed issue where "selectin_polymorphic" loading for inheritance mappers
would not function correctly if the :paramref:`_orm.Mapper.polymorphic_on`
parameter referred to a SQL expression that was not directly mapped on the
class.

.. change::
:tags: bug, orm
:tickets: 8710
:versions: 2.0.0b3

Fixed issue where the underlying DBAPI cursor would not be closed when
using the :class:`_orm.Query` object as an iterator, if a user-defined exception
case were raised within the iteration process, thereby causing the iterator
to be closed by the Python interpreter. When using
:meth:`_orm.Query.yield_per` to create server-side cursors, this would lead
to the usual MySQL-related issues with server side cursors out of sync,
and without direct access to the :class:`.Result` object, end-user code
could not access the cursor in order to close it.

To resolve, a catch for ``GeneratorExit`` is applied within the iterator
method, which will close the result object in those cases when the
iterator were interrupted, and by definition will be closed by the
Python interpreter.

As part of this change as implemented for the 1.4 series, ensured that
``.close()`` methods are available on all :class:`.Result` implementations
including :class:`.ScalarResult`, :class:`.MappingResult`. The 2.0
version of this change also includes new context manager patterns for use
with :class:`.Result` classes.

.. change::
:tags: bug, engine
:tickets: 8710

Ensured all :class:`.Result` objects include a :meth:`.Result.close` method
as well as a :attr:`.Result.closed` attribute, including on
:class:`.ScalarResult` and :class:`.MappingResult`.

.. change::
:tags: bug, mssql, reflection
:tickets: 8700
:versions: 2.0.0b3

Fixed issue with :meth:`.Inspector.has_table`, which when used against a
view with the SQL Server dialect would erroneously return ``False``, due to
a regression in the 1.4 series which removed support for this on SQL
Server. The issue is not present in the 2.0 series which uses a different
reflection architecture. Test support is added to ensure ``has_table()``
remains working per spec re: views.

.. change::
:tags: bug, sql
:tickets: 8724
:versions: 2.0.0b3

Fixed issue which prevented the :func:`_sql.literal_column` construct from
working properly within the context of a :class:`.Select` construct as well
as other potential places where "anonymized labels" might be generated, if
the literal expression contained characters which could interfere with
format strings, such as open parenthesis, due to an implementation detail
of the "anonymous label" structure.


.. changelog::

:released: October 16, 2022

.. change::
:tags: bug, asyncio
:tickets: 8516

Improved implementation of ``asyncio.shield()`` used in context managers as
added in :ticket:`8145`, such that the "close" operation is enclosed within
an ``asyncio.Task`` which is then strongly referenced as the operation
proceeds. This is per Python documentation indicating that the task is
otherwise not strongly referenced.

.. change::
:tags: bug, orm
:tickets: 8614

The :paramref:`_orm.Session.execute.bind_arguments` dictionary is no longer
mutated when passed to :meth:`_orm.Session.execute` and similar; instead,
it's copied to an internal dictionary for state changes. Among other
things, this fixes and issue where the "clause" passed to the
:meth:`_orm.Session.get_bind` method would be incorrectly referring to the
:class:`_sql.Select` construct used for the "fetch" synchronization
strategy, when the actual query being emitted was a :class:`_dml.Delete` or
:class:`_dml.Update`. This would interfere with recipes for "routing
sessions".

.. change::
:tags: bug, orm
:tickets: 7094

A warning is emitted in ORM configurations when an explicit
:func:`_orm.remote` annotation is applied to columns that are local to the
immediate mapped class, when the referenced class does not include any of
the same table columns. Ideally this would raise an error at some point as
it's not correct from a mapping point of view.

.. change::
:tags: bug, orm
:tickets: 7545

A warning is emitted when attempting to configure a mapped class within an
inheritance hierarchy where the mapper is not given any polymorphic
identity, however there is a polymorphic discriminator column assigned.
Such classes should be abstract if they never intend to load directly.


.. change::
:tags: bug, mssql, regression
:tickets: 8525

Fixed yet another regression in SQL Server isolation level fetch (see
:ticket:`8231`, :ticket:`8475`), this time with "Microsoft Dynamics CRM
Database via Azure Active Directory", which apparently lacks the
``system_views`` view entirely. Error catching has been extended that under
no circumstances will this method ever fail, provided database connectivity
is present.

.. change::
:tags: orm, bug, regression
:tickets: 8569

Fixed regression for 1.4 in :func:`_orm.contains_eager` where the "wrap in
subquery" logic of :func:`_orm.joinedload` would be inadvertently triggered
for use of the :func:`_orm.contains_eager` function with similar statements
(e.g. those that use ``distinct()``, ``limit()`` or ``offset()``), which
would then lead to secondary issues with queries that used some
combinations of SQL label names and aliasing. This "wrapping" is not
appropriate for :func:`_orm.contains_eager` which has always had the
contract that the user-defined SQL statement is unmodified with the
exception of adding the appropriate columns to be fetched.

.. change::
:tags: bug, orm, regression
:tickets: 8507

Fixed regression where using ORM update() with synchronize_session='fetch'
would fail due to the use of evaluators that are now used to determine the
in-Python value for expressions in the SET clause when refreshing
objects; if the evaluators make use of math operators against non-numeric
values such as PostgreSQL JSONB, the non-evaluable condition would fail to
be detected correctly. The evaluator now limits the use of math mutation
operators to numeric types only, with the exception of "+" that continues
to work for strings as well. SQLAlchemy 2.0 may alter this further by
fetching the SET values completely rather than using evaluation.

.. change::
:tags: usecase, postgresql
:tickets: 8574

:class:`_postgresql.aggregate_order_by` now supports cache generation.

.. change::
:tags: bug, mysql
:tickets: 8588

Adjusted the regular expression used to match "CREATE VIEW" when
testing for views to work more flexibly, no longer requiring the
special keyword "ALGORITHM" in the middle, which was intended to be
optional but was not working correctly. The change allows view reflection
to work more completely on MySQL-compatible variants such as StarRocks.
Pull request courtesy John Bodley.

.. change::
:tags: bug, engine
:tickets: 8536

Fixed issue where mixing "*" with additional explicitly-named column
expressions within the columns clause of a :func:`_sql.select` construct
would cause result-column targeting to sometimes consider the label name or
other non-repeated names to be an ambiguous target.

.. changelog::

:released: September 6, 2022

.. change::
:tags: bug, sql
:tickets: 8441

Fixed issue where use of the :func:`_sql.table` construct, passing a string
for the :paramref:`_sql.table.schema` parameter, would fail to take the
"schema" string into account when producing a cache key, thus leading to
caching collisions if multiple, same-named :func:`_sql.table` constructs
with different schemas were used.


.. change::
:tags: bug, events, orm
:tickets: 8467

Fixed event listening issue where event listeners added to a superclass
would be lost if a subclass were created which then had its own listeners
associated. The practical example is that of the :class:`.sessionmaker`
class created after events have been associated with the
:class:`_orm.Session` class.

.. change::
:tags: orm, bug
:tickets: 8401

Hardened the cache key strategy for the :func:`_orm.aliased` and
:func:`_orm.with_polymorphic` constructs. While no issue involving actual
statements being cached can easily be demonstrated (if at all), these two
constructs were not including enough of what makes them unique in their
cache keys for caching on the aliased construct alone to be accurate.

.. change::
:tags: bug, orm, regression
:tickets: 8456

Fixed regression appearing in the 1.4 series where a joined-inheritance
query placed as a subquery within an enclosing query for that same entity
would fail to render the JOIN correctly for the inner query. The issue
manifested in two different ways prior and subsequent to version 1.4.18
(related issue :ticket:`6595`), in one case rendering JOIN twice, in the
other losing the JOIN entirely. To resolve, the conditions under which
"polymorphic loading" are applied have been scaled back to not be invoked
for simple joined inheritance queries.

.. change::
:tags: bug, orm
:tickets: 8446

Fixed issue in :mod:`sqlalchemy.ext.mutable` extension where collection
links to the parent object would be lost if the object were merged with
:meth:`.Session.merge` while also passing :paramref:`.Session.merge.load`
as False.

.. change::
:tags: bug, orm
:tickets: 8399

Fixed issue involving :func:`_orm.with_loader_criteria` where a closure
variable used as bound parameter value within the lambda would not carry
forward correctly into additional relationship loaders such as
:func:`_orm.selectinload` and :func:`_orm.lazyload` after the statement
were cached, using the stale originally-cached value instead.


.. change::
:tags: bug, mssql, regression
:tickets: 8475

Fixed regression caused by the fix for :ticket:`8231` released in 1.4.40
where connection would fail if the user did not have permission to query
the ``dm_exec_sessions`` or ``dm_pdw_nodes_exec_sessions`` system views
when trying to determine the current transaction isolation level.

.. change::
:tags: bug, asyncio
:tickets: 8419

Integrated support for asyncpg's ``terminate()`` method call for cases
where the connection pool is recycling a possibly timed-out connection,
where a connection is being garbage collected that wasn't gracefully
closed, as well as when the connection has been invalidated. This allows
asyncpg to abandon the connection without waiting for a response that may
incur long timeouts.

.. changelog::

:released: August 8, 2022

.. change::
:tags: bug, orm
:tickets: 8357

Fixed issue where referencing a CTE multiple times in conjunction with a
polymorphic SELECT could result in multiple "clones" of the same CTE being
constructed, which would then trigger these two CTEs as duplicates. To
resolve, the two CTEs are deep-compared when this occurs to ensure that
they are equivalent, then are treated as equivalent.


.. change::
:tags: bug, orm, declarative
:tickets: 8190

Fixed issue where a hierarchy of classes set up as an abstract or mixin
declarative classes could not declare standalone columns on a superclass
that would then be copied correctly to a :class:`_orm.declared_attr`
callable that wanted to make use of them on a descendant class.

.. change::
:tags: bug, types
:tickets: 7249

Fixed issue where :class:`.TypeDecorator` would not correctly proxy the
``__getitem__()`` operator when decorating the :class:`_types.ARRAY`
datatype, without explicit workarounds.

.. change::
:tags: bug, asyncio
:tickets: 8145

Added ``asyncio.shield()`` to the connection and session release process
specifically within the ``__aexit__()`` context manager exit, when using
:class:`.AsyncConnection` or :class:`.AsyncSession` as a context manager
that releases the object when the context manager is complete. This appears
to help with task cancellation when using alternate concurrency libraries
such as ``anyio``, ``uvloop`` that otherwise don't provide an async context
for the connection pool to release the connection properly during task
cancellation.



.. change::
:tags: bug, postgresql
:tickets: 4392

Fixed issue in psycopg2 dialect where the "multiple hosts" feature
implemented for :ticket:`4392`, where multiple ``host:port`` pairs could be
passed in the query string as
``?host=host1:port1&host=host2:port2&host=host3:port3`` was not implemented
correctly, as it did not propagate the "port" parameter appropriately.
Connections that didn't use a different "port" likely worked without issue,
and connections that had "port" for some of the entries may have
incorrectly passed on that hostname. The format is now corrected to pass
hosts/ports appropriately.

As part of this change, maintained support for another multihost style that
worked unintentionally, which is comma-separated
``?host=h1,h2,h3&port=p1,p2,p3``. This format is more consistent with
libpq's query-string format, whereas the previous format is inspired by a
different aspect of libpq's URI format but is not quite the same thing.

If the two styles are mixed together, an error is raised as this is
ambiguous.

.. change::
:tags: bug, sql
:tickets: 8253

Adjusted the SQL compilation for string containment functions
``.contains()``, ``.startswith()``, ``.endswith()`` to force the use of the
string concatenation operator, rather than relying upon the overload of the
addition operator, so that non-standard use of these operators with for
example bytestrings still produces string concatenation operators.


.. change::
:tags: bug, orm
:tickets: 8235

A :func:`_sql.select` construct that is passed a sole '*' argument for
``SELECT *``, either via string, :func:`_sql.text`, or
:func:`_sql.literal_column`, will be interpreted as a Core-level SQL
statement rather than as an ORM level statement. This is so that the ``*``,
when expanded to match any number of columns, will result in all columns
returned in the result. the ORM- level interpretation of
:func:`_sql.select` needs to know the names and types of all ORM columns up
front which can't be achieved when ``'*'`` is used.

If ``'*`` is used amongst other expressions simultaneously with an ORM
statement, an error is raised as this can't be interpreted correctly by the
ORM.

.. change::
:tags: bug, mssql
:tickets: 8210

Fixed issues that prevented the new usage patterns for using DML with ORM
objects presented at :ref:`orm_dml_returning_objects` from working
correctly with the SQL Server pyodbc dialect.


.. change::
:tags: bug, mssql
:tickets: 8231

Fixed issue where the SQL Server dialect's query for the current isolation
level would fail on Azure Synapse Analytics, due to the way in which this
database handles transaction rollbacks after an error has occurred. The
initial query has been modified to no longer rely upon catching an error
when attempting to detect the appropriate system view. Additionally, to
better support this database's very specific "rollback" behavior,
implemented new parameter ``ignore_no_transaction_on_rollback`` indicating
that a rollback should ignore Azure Synapse error 'No corresponding
transaction found. (111214)', which is raised if no transaction is present
in conflict with the Python DBAPI.

Initial patch and valuable debugging assistance courtesy of ww2406.

.. seealso::

:ref:`azure_synapse_ignore_no_transaction_on_rollback`

.. change::
:tags: bug, mypy
:tickets: 8196

Fixed a crash of the mypy plugin when using a lambda as a Column
default. Pull request courtesy of tchapi.


.. change::
:tags: usecase, engine

Implemented new :paramref:`_engine.Connection.execution_options.yield_per`
execution option for :class:`_engine.Connection` in Core, to mirror that of
the same :ref:`yield_per <orm_queryguide_yield_per>` option available in
the ORM. The option sets both the
:paramref:`_engine.Connection.execution_options.stream_results` option at
the same time as invoking :meth:`_engine.Result.yield_per`, to provide the
most common streaming result configuration which also mirrors that of the
ORM use case in its usage pattern.

.. seealso::

:ref:`engine_stream_results` - revised documentation


.. change::
:tags: bug, engine

Fixed bug in :class:`_engine.Result` where the usage of a buffered result
strategy would not be used if the dialect in use did not support an
explicit "server side cursor" setting, when using
:paramref:`_engine.Connection.execution_options.stream_results`. This is in
error as DBAPIs such as that of SQLite and Oracle already use a
non-buffered result fetching scheme, which still benefits from usage of
partial result fetching. The "buffered" strategy is now used in all
cases where :paramref:`_engine.Connection.execution_options.stream_results`
is set.


.. change::
:tags: bug, engine
:tickets: 8199

Added :meth:`.FilterResult.yield_per` so that result implementations
such as :class:`.MappingResult`, :class:`.ScalarResult` and
:class:`.AsyncResult` have access to this method.

.. changelog::

:released: June 24, 2022

.. change::
:tags: bug, orm, regression
:tickets: 8133

Fixed regression caused by :ticket:`8133` where the pickle format for
mutable attributes was changed, without a fallback to recognize the old
format, causing in-place upgrades of SQLAlchemy to no longer be able to
read pickled data from previous versions. A check plus a fallback for the
old format is now in place.

.. changelog::