fix(python): Assert SQLInterfaceError is raised

[wence-]

No description provided.

[wence-]

cc @alexander-beedie

[MarcoGorelli]

thanks @wence- !

[MarcoGorelli]

aah looks like there's another issue in that file

_________________________ test_frame_sql_globals_error _________________________
[gw0] linux -- Python 3.8.18 /home/runner/work/polars/polars/py-polars/.venv/bin/python3
tests/unit/sql/test_miscellaneous.py:99: in test_frame_sql_globals_error
    df1.sql(query=query)
polars/dataframe/frame.py:4734: in sql
    return ctx.execute(query)  # type: ignore[return-value]
polars/sql/context.py:281: in execute
    res = wrap_ldf(self._ctxt.execute(query))
E   polars.exceptions.SQLInterfaceError: relation 'df2' was not found

[alexander-beedie]

I believe you just need to rebase; there are now dedicated SQLInterface & SQLSyntax errors.

[wence-]

aah looks like there's another issue in that file

Thanks, I didn't actually run this, which was a mistake

[wence-]

I believe you just need to rebase; there are now dedicated SQLInterface & SQLSyntax errors.

We need to actually catch SQLInterfaceError in this case though, right? (As @MarcoGorelli just pushed) Rather than ComputeError as is in main.

[MarcoGorelli]

There's a hypothesis test that has failed, but that's unrelated to what you're fixing

FAILED tests/unit/operations/namespaces/temporal/test_datetime.py::test_series_datetime_timeunits - polars.exceptions.ComputeError: datetime '1959-10-25 01:00:00' is ambiguous in time zone 'America/Nassau'. Please use `ambiguous` to tell how it should be localized.
while generating 's' from series(dtype=Datetime, min_size=1, max_size=10, allow_null=False)
You can reproduce this example by temporarily adding @reproduce_failure('6.103.0', b'AAABAKkAAQApCRcUAAAAAAAA') as a decorator on your test case

I think this can be merged to get CI green again, I'll make a separate PR to address the hypothesis failure

[alexander-beedie]

Ahhh, I see what happened; merging the backlog of breaking changes left this test a little out of order. It's already fixed in #16711, for example, which is from a branch that was pending merge of the breaking changes and accounts for them. Can merge either to get this test green again.

[MarcoGorelli]

if ##16711 already fixes it, let's go with that?

closing this one then, but thanks @wence- for your PR!

[MarcoGorelli]

@wence- are you still able to restore this branch? it may take a little longer to get #16711 in, if you could restore it and reopen I think we could just merge and get ci green again

[wence-]

@wence- are you still able to restore this branch? it may take a little longer to get #16711 in, if you could restore it and reopen I think we could just merge and get ci green again

Done (restored), I don't have permissions to reopen this PR though.

[MarcoGorelli]

thanks!

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.47%. Comparing base (9912af0) to head (d57da14).
Report is 16 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16713      +/-   ##
==========================================
- Coverage   81.49%   81.47%   -0.02%     
==========================================
  Files        1416     1415       -1     
  Lines      186880   186768     -112     
  Branches     3023     3022       -1     
==========================================
- Hits       152289   152167     -122     
- Misses      34059    34069      +10     
  Partials      532      532              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[MarcoGorelli]

thanks @wence- !