Don't check order for filtered groupby test

[charlesbluca]

Looks like changes in Dask have made the ordering of one of the results in test_group_by_filtered different, causing failures - this PR adds check_index=False to the relevant assertion, since maintaining a sorted ordering for groupbys wasn't a guarantee to begin with.

Perhaps it would be worth thinking about if we want to make this a default kwarg for our assert_eq shim, considering we aren't concerned with ordering unless something like an ORDER BY statement is being tested (cc @ayushdg).

Closes #701

[ayushdg]

Perhaps it would be worth thinking about if we want to make this a default kwarg for our assert_eq shim, considering we aren't concerned with ordering unless something like an ORDER BY statement is being tested (cc @ayushdg).

Thanks for triaging. I agree that it makes sense to set this to false for all groupby (and maybe join) related tests. But I do think for standard rex calls and other tests we do want to ensure that input ordering is maintained (similar to sql). So it might make sense to limit this to groupby (and join) tests for now.

Unrelated but gpuCi failures will probably be fixed via rapidsai/cudf#11561

[codecov-commenter]

Codecov Report

Merging #702 (a90beec) into main (c9b8e2e) will increase coverage by 2.08%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #702      +/-   ##
==========================================
+ Coverage   86.53%   88.62%   +2.08%     
==========================================
  Files          74       69       -5     
  Lines        4671     3507    -1164     
  Branches     1090      710     -380     
==========================================
- Hits         4042     3108     -934     
+ Misses        491      308     -183     
+ Partials      138       91      -47     

Impacted Files	Coverage Δ
dask_sql/physical/rex/base.py	77.77% <0.00%> (-3.48%)	⬇️
dask_sql/physical/rel/custom/drop_table.py	82.35% <0.00%> (-3.37%)	⬇️
dask_sql/physical/rel/base.py	92.10% <0.00%> (-2.73%)	⬇️
dask_sql/physical/rel/logical/union.py	91.17% <0.00%> (-2.71%)	⬇️
dask_sql/physical/rex/core/input_ref.py	80.00% <0.00%> (-2.36%)	⬇️
dask_sql/physical/rel/logical/filter.py	84.37% <0.00%> (-2.12%)	⬇️
dask_sql/physical/rex/convert.py	88.00% <0.00%> (-1.75%)	⬇️
dask_sql/physical/rel/logical/limit.py	92.85% <0.00%> (-1.27%)	⬇️
dask_sql/physical/rel/convert.py	87.50% <0.00%> (-0.74%)	⬇️
dask_sql/_version.py	34.00% <0.00%> (-0.29%)	⬇️
... and 23 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[charlesbluca]

Going to merge this for now to unblock CI, we can think about relaxing the assert_eq checks for specific tests in a follow up