fix length error with array_has

[samuelcolvin]

Rationale for this change

Fixes an error we've been experiencing in prod where some queries that use array_has (including via ANY()) result int:

DataFusion CLI v41.0.0
Error: Internal error: UDF returned a different number of rows than expected. Expected: 8192, Got: 1.
This was likely caused by a bug in DataFusion's code and we would welcome that you file an bug report in our issue tracker

What changes are included in this PR?

Fix a bug in array_has where it returned 1 row, instead of the size of the group.

Are these changes tested?

Yes.

Are there any user-facing changes?

No.

[samuelcolvin]

@alamb I would love to get this merged, it's causing many queries that use ANY() to fail right now.

cc @jayzhan211 who I think wrote this.

[samuelcolvin]

the point is that in this case the array returned should be the same length as the haystack, not 1.

[alamb]

Thank you for this fix @samuelcolvin -- I have one small nit pick about where the test should live, but we can also fix that as a follow on PR.

[samuelcolvin]

slt test updated and dedicated test removed.

It would be great to get this merged. We'll run datafusion main until there's a another release since this is is affecting a lot of users as we migrate querying from timescale to datafusion.

[alamb]

Thanks @jayzhan211 and @samuelcolvin

[alamb]

beautiful 🎉

[samuelcolvin]

Thanks @alamb for merging so quickly. 🙏

[alamb]

Thanks @alamb for merging so quickly. 🙏

I try hard to prioritize bug fixes. Thank you for fixing the bug!