test: re-enable window function over parquet with forced collisions

[korowa]

Which issue does this PR close?

Closes#11660.

Rationale for this change

After several attempts to reproduce the failure locally and in fork CI it still was not reproducing. Also I wasn't able to find any workflow failure related to 1.80 upgrade in PRs and on main. So, trying to just uncomment the test.

What changes are included in this PR?

Test uncommented.

Are these changes tested?

Are there any user-facing changes?

[alamb]

Thank you @korowa

FYI @nrc who originally reported this was failing

[alamb]

I re-ran the relevant test:

cargo test --test sqllogictests --features=force_hash_collisions -- parquet

And indeed it passes locally for me too

andrewlamb@Andrews-MBP-2:~/Software/datafusion$ cargo test --test sqllogictests --features=force_hash_collisions -- parquet
warning: function `hash_map_array` is never used
--> datafusion/common/src/hash_utils.rs:249:4
|
249 | fn hash_map_array(
| ^^^^^^^^^^^^^^
|
= note: `#[warn(dead_code)]` on by default
warning: `datafusion-common` (lib) generated 1 warning
Finished `test` profile [unoptimized + debuginfo] target(s) in 0.40s
Running bin/sqllogictests.rs (target/debug/deps/sqllogictests-a3320b7543efea67)
Running "parquet.slt"
Running "parquet_sorted_statistics.slt"
andrewlamb@Andrews-MBP-2:~/Software/datafusion$

[comphead]

Suggested change

# Perform the query:

[comphead]

Would be great to add some comments why test is so specific

[alamb]

I agree better comments would be nice -- however I don't think this PR makes the comment any more/less cryptic, so I don't think we should hold it for better comments.

Maybe we can improve the comments as a future PR

[alamb]

Thanks agian @korowa and @comphead

[alamb]

I agree better comments would be nice -- however I don't think this PR makes the comment any more/less cryptic, so I don't think we should hold it for better comments.

Maybe we can improve the comments as a future PR