Mike Bayer [Fri, 1 Aug 2025 16:48:15 +0000 (12:48 -0400)]
Fix PostgreSQL JSONB subscripting regression with functions
Fixed regression in PostgreSQL dialect where JSONB subscription syntax
would generate incorrect SQL for JSONB-returning functions, causing syntax
errors. The dialect now properly wraps function calls and expressions in
parentheses when using the ``[]`` subscription syntax, generating
``(function_call)[index]`` instead of ``function_call[index]`` to comply
with PostgreSQL syntax requirements.
Added ``dataclass_metadata`` argument to a all column functions
used in the ORM that accept dataclasses parameters.
It's passed to the underlying dataclass ``metadata`` attribute
of the dataclass field.
Pull request courtesy Sigmund Lahn.
Mike Bayer [Mon, 14 Jul 2025 20:13:03 +0000 (16:13 -0400)]
extend Values from HasCTE
The :func:`_sql.values` construct gains a new method :meth:`_sql.Values.cte`,
which allows creation of a named, explicit-columns :class:`.CTE` against an
unnamed ``VALUES`` expression, producing a syntax that allows column-oriented
selection from a ``VALUES`` construct on modern versions of PostgreSQL, SQLite,
and MariaDB.
Federico Caselli [Fri, 29 Dec 2023 21:30:24 +0000 (22:30 +0100)]
Improve handling of sentinel columns
Fixed issue in "insertmanyvalues" feature where an INSERT..RETURNING
that also made use of a sentinel column to track results would fail to
filter out the additional column when :meth:`.Result.unique` were used
to uniquify the result set.
Mike Bayer [Thu, 10 Jul 2025 19:29:29 +0000 (15:29 -0400)]
Unconditionally flush session on all statement executions
This change simplifies session execution logic to always run autoflush
for all statement executions, regardless of whether they are ORM or Core
statements. Previously, autoflush was conditionally applied only when
ORM-related statements were detected, which had become difficult to define
clearly with the unified v2 syntax that blurs the lines between Core and
ORM execution patterns.
The implementation adds a call to self._autoflush() in the Core execution
path (when compile_state_cls is None) in Session._execute_internal(),
ensuring consistent autoflush behavior across all types of SQL execution.
This provides more predictable session behavior and eliminates the
confusion around which statements trigger autoflush.
Also included is some improvement to the migration_21 document
Mike Bayer [Fri, 11 Jul 2025 16:51:44 +0000 (12:51 -0400)]
support JSONB subscripting syntax
Added support for PostgreSQL 14+ JSONB subscripting syntax. When connected
to PostgreSQL 14 or later, JSONB columns now automatically use the native
subscript notation ``jsonb_col['key']`` instead of the arrow operator
``jsonb_col -> 'key'`` for both read and write operations. This provides
better compatibility with PostgreSQL's native JSONB subscripting feature
while maintaining backward compatibility with older PostgreSQL versions.
JSON columns continue to use the traditional arrow syntax regardless of
PostgreSQL version.
Mike Bayer [Thu, 3 Apr 2025 14:36:28 +0000 (10:36 -0400)]
generalize composite bulk insert to hybrids
Added new hybrid method :meth:`.hybrid_property.bulk_insert_setter` which
works in a similar way as :meth:`.hybrid_property.update_expression` for
bulk ORM operations. A user-defined class method can now populate a bulk
insert mapping dictionary using the desired hybrid mechanics. New
documentation is added showing how both of these methods can be used
including in combination with the new :func:`_sql.from_dml_column`
construct.
Fix the code examples and the grammatic in `declarative_tables.rst` (#12721)
* Fix the code example in `declarative_tables.rst`
Add an import of a `Optional` class from the `typing` library in the
code example of the "Union types inside the Type Map" chapter.
* Fix the code example in `declarative_tables.rst`
Replace the import of the `deferred` function to `column_property` from
the `sqlalchemy.orm` package in first code example of the "Applying
Load, Persistence and Mapping Options for Imperative Table Columns"
chapter.
* Fix the grammatic in `declarative_tables.rst`
Remove the unnecessary article in the second paragraph of the
"Mapping to an Explicit Set of Primary Key Columns" chapter in
`declarative_tables.rst`.
Mike Bayer [Thu, 3 Apr 2025 14:36:28 +0000 (10:36 -0400)]
add CRUD column marker
Added new Core feature :func:`_sql.from_dml_column` that may be used in
expressions inside of :meth:`.UpdateBase.values` for INSERT or UPDATE; this
construct will copy whatever SQL expression is used for the given target
column in the statement to be used with additional columns. The construct
is mostly intended to be a helper with ORM :class:`.hybrid_property` within
DML hooks.
This is the Core side of the feature being added to the ORM
for #12496
Mike Bayer [Thu, 19 Jun 2025 16:39:17 +0000 (12:39 -0400)]
support insert of table columns in specific positions
Added method :meth:`.TableClause.insert_column` to complement
:meth:`.TableClause.append_column`, which inserts the given column at a
specific index. This can be helpful for prepending primary key columns to
tables, etc.
Federico Caselli [Fri, 20 Jun 2025 20:28:45 +0000 (22:28 +0200)]
Add new str` subclass for postgresql bitstring
Adds a new ``str`` subclass :class:`dialects.postgresql.BitString`
representing PostgreSQL bitstrings in python, that includes
functionality for converting to and from ``int`` and ``bytes``, in
addition to implementing utility methods and operators for dealing
with bits.
This new class is returned automatically by the :class:`postgresql.BIT`
type.
remove support for list of tuples in the normal execute
The function that validates the arguments in the normal execute flow
allowed by mistake list of tuples, that are not supported by the
code since the 2.0 series.
Mike Bayer [Tue, 24 Jun 2025 17:57:17 +0000 (13:57 -0400)]
try flake8-import-order 0.19.2
the big new thang is that it is doing import order checks inside of
TYPE_CHECKING blocks. Introduces some new codes that we
enthusiastically add to our ignore list. update to the latest
and greatest zimports 0.6.2 as well
Mike Bayer [Mon, 23 Jun 2025 13:21:59 +0000 (09:21 -0400)]
hardening against inappropriate multi-table updates
Hardening of the compiler's actions for UPDATE statements that access
multiple tables to report more specifically when tables or aliases are
referenced in the SET clause; on cases where the backend does not support
secondary tables in the SET clause, an explicit error is raised, and on the
MySQL or similar backends that support such a SET clause, more specific
checking for not-properly-included tables is performed. Overall the change
is preventing these erroneous forms of UPDATE statements from being
compiled, whereas previously it was relied on the database to raise an
error, which was not always guaranteed to happen, or to be non-ambiguous,
due to cases where the parent table included the same column name as the
secondary table column being updated.
Fixed bug where the ORM would pull in the wrong column into an UPDATE when
a key name inside of the :meth:`.ValuesBase.values` method could be located
from an ORM entity mentioned in the statement, but where that ORM entity
was not the actual table that the statement was inserting or updating. An
extra check for this edge case is added to avoid this problem.
Mike Fiedler [Fri, 20 Jun 2025 21:49:26 +0000 (17:49 -0400)]
Update `table_per_related` example for Declarative API
<!-- Provide a general summary of your proposed changes in the Title field above -->
### Description
When reading more of the Generic Associations, I found that the examples state "Annotated example", and wondered what that meant, since in other parts of the docs that means they are using the new 2.0 style.
I tried to update this example to be more in line with the new style, including a little f-string update.
I completely understand this is unlikely to be merged as-is - but wanted to understand more about the right way to use modern styles to properly code well-hinted, more "exotic" implementations.
Outstanding questions:
- Should examples pass pass `mypy --strict` ? It doesn't right now.
- Are there better ways to apply `Mapped` within the `type(...)` definition, so we could skip importing `Integer`?
### Checklist
<!-- go over following points. check them with an `x` if they do apply, (they turn into clickable checkboxes once the PR is submitted, so no need to do everything at once)
-->
This pull request is:
(Do non-running examples count as code fixes/features?)
- [x] A documentation / typographical / small typing error fix
- Good to go, no issue or tests are needed
Mike Bayer [Fri, 20 Jun 2025 17:04:54 +0000 (13:04 -0400)]
really pin flake8-import-order to <0.19
In 8a287bf5c5635daf99217e I told @caselit to just omit 0.19.0,
as I wa sure 0.19.1 would fix the issue. Unfortunately in [1]
we see that this release continues to have problems, which seems
to be related not to the importlib.metadata fix but to an unrelated
improvement for TYPE_CHECKING blocks.
So we really have to wait for them to get a working release before
we can update.
Mike Bayer [Mon, 16 Jun 2025 23:53:30 +0000 (19:53 -0400)]
rework wraps_column_expression logic to be purely compile time checking
Fixed issue where :func:`.select` of a free-standing, unnamed scalar expression that
has a unary operator applied, such as negation, would not apply result
processors to the selected column even though the correct type remains in
place for the unary expression.
This change opened up a typing rabbithole where we were led to also
improve and harden the typing for the Exists element, in particular
in that the Exists now always refers to a ScalarSelect object, and
no longer a SelectStatementGrouping within the _regroup() cases; there
did not seem to be any reason for this inconsistency.
Mike Bayer [Wed, 11 Jun 2025 18:55:14 +0000 (14:55 -0400)]
update pickle tests
Since I want to get rid of util.portable_instancemethod, first
make sure we are testing pickle extensively including going through
all protocols for all metadata-oriented tests.
Mike Bayer [Tue, 10 Jun 2025 18:51:57 +0000 (14:51 -0400)]
remove util.portable_instancemethod
python seems to be able to pickle instance methods since
version 3.4. Doing a bisect shows it's
https://github.com/python/cpython/commit/c9dc4a2a8a6dcfe1674685bea4a4af935c0e37ca
where pickle protocol 4 was added, however we can see that protocols 0
through 4 also support pickling of methods. None of this documented.
Pablo Estevez [Mon, 9 Jun 2025 12:49:13 +0000 (08:49 -0400)]
update tox mypy
<!-- Provide a general summary of your proposed changes in the Title field above -->
After this commit https://github.com/sqlalchemy/sqlalchemy/commit/68cd3e8ec7098d4bb4b2102ad247f84cd89dfd8c
tox will fail with mypy below 1.16, at least locally.
### Description
<!-- Describe your changes in detail -->
### Checklist
<!-- go over following points. check them with an `x` if they do apply, (they turn into clickable checkboxes once the PR is submitted, so no need to do everything at once)
-->
This pull request is:
- [ ] A documentation / typographical / small typing error fix
- Good to go, no issue or tests are needed
- [ ] A short code fix
- please include the issue number, and create an issue if none exists, which
must include a complete example of the issue. one line code fixes without an
issue and demonstration will not be accepted.
- Please include: `Fixes: #<issue number>` in the commit message
- please include tests. one line code fixes without tests will not be accepted.
- [ ] A new feature implementation
- please include the issue number, and create an issue if none exists, which must
include a complete example of how the feature would look.
- Please include: `Fixes: #<issue number>` in the commit message
- please include tests.
Mike Bayer [Sun, 8 Jun 2025 17:01:45 +0000 (13:01 -0400)]
use sys.columns to allow accurate joining to other SYS tables
Reworked SQL Server column reflection to be based on the ``sys.columns``
table rather than ``information_schema.columns`` view. By correctly using
the SQL Server ``object_id()`` function as a lead and joining to related
tables on object_id rather than names, this repairs a variety of issues in
SQL Server reflection, including:
* Issue where reflected column comments would not correctly line up
with the columns themselves in the case that the table had been ALTERed
* Correctly targets tables with awkward names such as names with brackets,
when reflecting not just the basic table / columns but also extended
information including IDENTITY, computed columns, comments which
did not work previously
* Correctly targets IDENTITY, computed status from temporary tables
which did not work previously
Mike Bayer [Thu, 5 Jun 2025 12:58:49 +0000 (08:58 -0400)]
hardcode now(), current_timstamp() into the MySQL regex
Fixed yet another regression caused by by the DEFAULT rendering changes in
2.0.40 :ticket:`12425`, similar to :ticket:`12488`, this time where using a
CURRENT_TIMESTAMP function with a fractional seconds portion inside a
textual default value would also fail to be recognized as a
non-parenthesized server default.
There's no way to do this other than start hardcoding a list
of MySQL functions that demand that parenthesis are not added around
them, I can think of no other heuristic that will work here.
Suggestions welcome
Mike Bayer [Tue, 3 Jun 2025 21:15:54 +0000 (17:15 -0400)]
use exact py3.14 version
gh actions is not complaining that the exact string "3.13", "3.12"
etc are not in versions-manifest.json, but for 3.14 it's complaining.
not happy to hardcode this but just to get it running
Federico Caselli [Wed, 28 May 2025 20:03:51 +0000 (22:03 +0200)]
Simplify postgresql index reflection query
Match on python side the values of `pg_am` and `pg_opclass`
to avoid joining them in the main query.
Since both queries have a limited size and are generally
stable their value can be cached using the inspector
cache.
Denis Laxalde [Wed, 28 May 2025 19:37:36 +0000 (15:37 -0400)]
Reflect index's column operator class on PostgreSQL
Fill the `postgresql_ops` key of PostgreSQL's `dialect_options` returned by get_multi_indexes() with a mapping from column names to the operator class, if it's not the default for respective data type.
As we need to join on ``pg_catalog.pg_opclass``, the table definition is added to ``postgresql.pg_catalog``.
Federico Caselli [Sun, 23 Jun 2024 13:01:40 +0000 (15:01 +0200)]
The ``Enum.inherit_schema`` now defaults to true
Changed the default value of :paramref:`_types.Enum.inherit_schema` to
``True`` when :paramref:`_types.Enum.schema` and
:paramref:`_types.Enum.metadata` parameters are not provided.
The same behavior has been applied also to PostgreSQL
:class:`_postgresql.DOMAIN` type.
projects / thirdparty / sqlalchemy / sqlalchemy.git / log
