agronholm / sqlacodegen

98%

LAST BUILD ON BRANCH master

branch: SELECT

CHANGE BRANCH

x

• No branch selected
• 2.0.0
• 2.0.1
• 2.1.0
• 2.2.0
• 2.3.0
• autoincrement-multiple-primary-keys
• better-field-name-for-relationships
• cli-schemas-arg-spaces
• dependabot
• dependabot/github_actions/dot-github/workflows/actions/download-artifact-4.1.7
• dependabot/github_actions/github-actions-7d67a68ff8
• dependabot/github_actions/github-actions-985357984d
• domain_check_textclause
• double-pluralization-of-names
• enum-generation-value-mapping
• feat.support_dialect
• feature/multi-file-output-pr
• filter-empty-dialect-args
• fix-285-char-collation
• fix-465
• fix-510
• fix-issue340-pgvector-genarating
• fix-mysql-double
• fix-sqlmodel-onetomany-multiref
• fix-tablename-pylance-types-and-add-missing-metadata-ref-for-sqlmodel-tables
• fix-typedecorator-adapt
• fix-usertypes-adapt
• fix/identity-decimal-serialization
• fix/mssql-sysname-reflection
• fix/postgresql-array-dialect-type
• fix/use-logging-instead-of-print
• maintainer
• master
• pre-commit-ci-update-config
• precommit-quarterly
• py314
• python-native-enums
• remove-engine-dependency
• render_column_type
• security-policy
• shared-name-import-fix
• sheinbergon/constraint-sa-version
• sheinbergon/domains-inherited-from-json
• sheinbergon/enum-generation-fixes
• sheinbergon/explicit_nullability_definition
• sheinbergon/not-implemented-python-type-handling
• sheinbergon/sqlalchemy-2_0_42-regression-fix
• spdx-license
• support-enum-arrays
• tidelift

Build # 27117669432

Build Type

push

github

Committed by web-flow

Commit Message

Preserve dialect-specific ARRAY types instead of adapting to generic ARRAY (#480)

When reflecting a PostgreSQL ``text[]`` (or any dialect-specific ARRAY) column, sqlacodegen previously walked the type's MRO in ``get_adapted_type`` and substituted ``sqlalchemy.dialects.postgresql.ARRAY`` with the generic ``sqlalchemy.ARRAY``. The generic ARRAY does not implement operators like ``.contains()``, ``.any()`` or ``.all()`` -- calling them on a generated model raises::

    NotImplementedError: ARRAY.contains() not implemented for the base ARRAY type; please use the dialect-specific ARRAY type

The workaround was to pass ``--options keep_dialect_types``, but that also forces every other type back to its dialect-specific form (``INTEGER`` instead of ``Integer``, etc.), which is more than the user wants.

This change special-cases ARRAY in ``get_adapted_type``: when the column type is a subclass of the generic ARRAY (i.e. a dialect-specific ARRAY), the original class is kept and only the item type is adapted. Plain generic ``sqlalchemy.ARRAY`` inputs are unaffected, and no other type's adaptation behavior changes.

The existing ``test_arrays`` is updated to expect the dialect-specific import, and a new ``test_array_preserves_dialect_for_runtime_operators`` test locks in the regression scenario described in the issue.

Fixes #441.

Coverage Stats

7 of 7 new or added lines in 2 files covered. (100.0%)

1 existing line in 1 file now uncovered.

1862 of 1904 relevant lines covered (97.79%)

4.89 hits per line