[ARROW-16548] [Python] Fix test_parquet_invalid_version test when no parquet or dataset are defined

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 8.0.0
• Fix Version/s: 9.0.0
• Component/s: Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/20241
• Assignee: Raúl Cumplido
• Reporter: Raúl Cumplido
• Created: 12/May/22 15:55
• Updated: 11/Jan/23 11:44
• Resolved: 17/May/22 09:58

Description

Working on the minimal_builds I have found that when building arrow and pyarrow without DATASET nor PARQUET the following error happens:

    def test_parquet_invalid_version(tempdir):
        table = pa.table({'a': [1, 2, 3]})
        with pytest.raises(ValueError, match="Unsupported Parquet format version"):
>           _write_table(table, tempdir / 'test_version.parquet', version="2.2")
E           NameError: name '_write_table' is not defined

pyarrow/tests/parquet/test_basic.py:52: NameError

This is the first test on the parquet/test_basic.py module.

I have done some investigation and if I change the test order (move that test on the file), the test that fails is a different one, always the first test on the module.
If I import (but not use)  from pyarrow.tests.parquet.common import pytestmark no test failures happen.

There seems to be a bug on how the pytestmark is injected. I am creating a ticket because it seems it might require more than a MINOR fix.

Issue Links

• links to: GitHub Pull Request #13147