feat: Support Utf8View for get_wider_type + binary_to_string_coercion functions

[jonathanc-n]

Which issue does this PR close?

Closes #13361
Closes #13360.

Rationale for this change

What changes are included in this PR?

Added string view to functions

Are these changes tested?

Are there any user-facing changes?

[jayzhan211]

Same with this, I would expect test that failed if the change is reverted

[jayzhan211]

Not this kind of test, but a valid sql query that goes through the coercion function

[jonathanc-n]

@jayzhan211 Correct me if I'm wrong, but this seems to be a problem with arrow casting from numerics to utf8view, the same can be said for this, #13366. However correct me if I'm wrong, I can't seem to find a sql query that triggers it, although i believe there should be one

[jayzhan211]

It seems only like_coercion utilize this coercion. I think current coercion is unnecessary complex for like operator

[Omega359]

I think get_wider_type is actually only used by ArrayConcat

[jayzhan211]

I mean binary_to_string_coercion

[Omega359]

Beyond the one bug I noted this looks fine to me. I do think however there should be a ticket filed to look at the get_wider_type function and see if it can be removed altogether as there seems to be just one use case and this may (I haven't checked but I feel it's likely) duplicate a function in arrow-cast.

[alamb]

I am trying to clean up the review backlog -- what is the status of this PR?

It seems like it is waiting on an end to end (e.g. sql or dataframe test) that actually has a problem / doesn't run without this PR but does run with it

I played around with it and I think I came up with one:

> create or replace table foo as select arrow_cast('1', 'Utf8View') as c;
0 row(s) fetched.
Elapsed 0.002 seconds.

> select c from foo;
+---+
| c |
+---+
| 1 |
+---+
1 row(s) fetched.
Elapsed 0.001 seconds.

> select c = 1 from foo;
Error during planning: Cannot infer common argument type for comparison operation Utf8View = Int64

But it works for strings:

> create or replace table foo as select arrow_cast('1', 'Utf8') as c;
0 row(s) fetched.
Elapsed 0.002 seconds.

> select c = 1 from foo;
+------------------+
| foo.c = Int64(1) |
+------------------+
| true             |
+------------------+
1 row(s) fetched.
Elapsed 0.002 seconds.

Looks like we need to test with Utf8 + Utf8View (aka string concat)

Also since the argument type order is important, also in both directions

Does that make sense @jonathanc-n ?

Thank you for your help

[jonathanc-n]

@alamb I also believe that this one is also waiting on numeric to utf8view coercion pr in arrow-rs, correct me if I'm wrong though.

[alamb]

We just updated to a version of arrow-rs that I think has the necessary changes:

• Upgrade arrow-rs, parquet to 54.0.0 and pyo3 to 0.23.3 #14153
  (thanks @Owen-CH-Leung )