]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/commit
use tuple expansion if type._is_tuple, test for Sequence if no type
authorMike Bayer <mike_mp@zzzcomputing.com>
Fri, 5 Nov 2021 14:18:42 +0000 (10:18 -0400)
committerMike Bayer <mike_mp@zzzcomputing.com>
Fri, 5 Nov 2021 15:55:14 +0000 (11:55 -0400)
commit0c44a1e77cfde0f841a4a64140314c6b833efdab
tree1678a98a75b4b40882c638b9193e52d3b931fd34
parent248d232459e38561999c4172acaaddd651c1a933
use tuple expansion if type._is_tuple, test for Sequence if no type

Fixed regression where the row objects returned for ORM queries, which are
now the normal :class:`_sql.Row` objects, would not be interpreted by the
:meth:`_sql.ColumnOperators.in_` operator as tuple values to be broken out
into individual bound parameters, and would instead pass them as single
values to the driver leading to failures. The change to the "expanding IN"
system now accommodates for the expression already being of type
:class:`.TupleType` and treats values accordingly if so. In the uncommon
case of using "tuple-in" with an untyped statement such as a textual
statement with no typing information, a tuple value is detected for values
that implement ``collections.abc.Sequence``, but that are not ``str`` or
``bytes``, as always when testing for ``Sequence``.

Added :class:`.TupleType` to the top level ``sqlalchemy`` import namespace.

Fixes: #7292
Change-Id: I8286387e3b3c3752b3bd4ae3560d4f31172acc22
doc/build/changelog/unreleased_14/7292.rst [new file with mode: 0644]
lib/sqlalchemy/__init__.py
lib/sqlalchemy/sql/compiler.py
lib/sqlalchemy/sql/sqltypes.py
lib/sqlalchemy/testing/suite/test_select.py
lib/sqlalchemy/types.py
lib/sqlalchemy/util/__init__.py
test/sql/test_lambdas.py
test/sql/test_resultset.py