Support Utf8View for string function bit_length()

[austin362667]

Which issue does this PR close?

Closes #13195

Rationale for this change

Thanks to @jayzhan211 , he noticed following issue which string scalar function bit_length() doesn't support Utf8View type:

select bit_length(arrow_cast('a', 'Utf8View'));

What changes are included in this PR?

Update bit_length() scalar function to support Utf8View

Are these changes tested?

Yes for scalar function. No for array function as it's still PR in apache/arrow-rs#6671

Are there any user-facing changes?

[alamb]

Thanks @austin362667 🙏

Can you please add a test to https://github.com/apache/datafusion/blob/main/datafusion/sqllogictest/test_files/string/string_query.slt.part ?

The structure is explained in https://github.com/apache/datafusion/blob/main/datafusion/sqllogictest/test_files/string/README.md

[austin362667]

certainly, thanks for review!

[alamb]

I don't understand this comment -- it seems like bit_length is working on Utf8View in datafusion/sqllogictest/test_files/string/string_view.slt

If you just uncomment this code out that would mean we are covering bit_length for all the string types

[austin362667]

Thanks @alamb
As I understand it, applying bit_length() to a column will currently fail because arrow::compute::kernels::length::bit_length does not yet support UTF-8 views (depends on PR in arrow-rs). However, it's still used in the following code path:

datafusion/datafusion/functions/src/string/bit_length.rs

Line 74 in a168548

ColumnarValue::Array(v) => Ok(ColumnarValue::Array(bit_length(v.as_ref())?)),

[goldmedal]

To cover other string types, we can add this test to each string testing file before apache/arrow-rs#6671.

[alamb]

Thank you @austin362667

[goldmedal]

I think we can move this test to string/string_literal.slt and then add similar tests for other string types as well (UTF8, LargeUTF8, and DictionaryString).

[goldmedal]

To cover other string types, we can add this test to each string testing file before apache/arrow-rs#6671.

[goldmedal]

I think we can add a TODO comment here and file an issue to track apache/arrow-rs#6671. After upgrading to the corresponding arrow-rs version, we should address the TODO comment in string/string_query.slt.part#L1100.

[alamb]

@austin362667 any chance you have some time to address @goldmedal 's suggestions?

[austin362667]

Thanks @goldmedal @alamb for the review. 🙏 Just addressed these issues
So sorry for the late reply.

[jayzhan211]

👍

[goldmedal]

Thanks @austin362667 and @alamb @jayzhan211 for reviewing. Although I think the tests can be improved, I can help to do that in the follow-up PR. Let's merge this PR first.