Support StringView for binary operators like `~`, `!~`, etc

[alamb]

Is your feature request related to a problem or challenge?

Part of https://github.com/apache/datafusion/issues/11752

As we work to complete StringView support in DataFusion @2010YOUY01 noticed on https://github.com/apache/datafusion/issues/11752#issuecomment-2308176932 that we don't currently support Regexp like binary operators https://datafusion.apache.org/user-guide/sql/operators.html#op-re-match for string view

Reproducer

CREATE TABLE t0(v0 DOUBLE, v1 DOUBLE, v2 BOOLEAN, v3 BOOLEAN, v4 BOOLEAN, v5 STRING);
INSERT INTO t0(v1, v5, v2) VALUES (0.7183242196192607, 'Tn', true);
CREATE TABLE t0_stringview AS SELECT v0, v1, v2, v3, v4, arrow_cast(v5, 'Utf8View') as v5 FROM t0;

> select v5 ~ 'foo' from t0_stringview;
Internal error: Data type Utf8View not supported for binary_string_array_flag_op_scalar operation 'regexp_is_match' on string array.
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

> select regexp_match(v5, 'foo') from t0_stringview;
+--------------------------------------------+
| regexp_match(t0_stringview.v5,Utf8("foo")) |
+--------------------------------------------+
|                                            |
+--------------------------------------------+
1 row(s) fetched.
Elapsed 0.034 seconds.

Describe the solution you'd like

StringView should be supported for these operators (aka the query should run without error)

Describe alternatives you've considered

Here are the relevant operator names:

            | Operator::RegexMatch
            | Operator::RegexIMatch
            | Operator::RegexNotMatch
            | Operator::RegexNotIMatch
            | Operator::LikeMatch
            | Operator::ILikeMatch
            | Operator::NotLikeMatch
            | Operator::NotILikeMatch

Here is the dispatch code:

https://github.com/apache/datafusion/blob/0f96af5b500efff72314f840a59a736787cc3def/datafusion/physical-expr/src/expressions/binary.rs#L621-L632

It appears that the corresponding arrow-rs kernel does not yet have support for StringView https://docs.rs/arrow-string/52.2.0/src/arrow_string/regexp.rs.html#307-311

So what I would suggest is:

1. Implement a PR in datafusion with coercion from Utf8View --> Utf8 (aka cast arguments back to string)
2. File an upstream ticket in arrow-rs for supporting string view with the regexp_like kernels and leave a link to that ticket in the datafusion code

Additional context

No response

[tlm365]

Looks like #12168 also might be a starting point to solve this problem. I also think about solution 2 when handling that PR, and IMO solution 2 is the better option. I will take this one.

File an upstream ticket in arrow-rs for supporting string view with the regexp_like kernels and leave a link to that ticket in the datafusion code

[tlm365]

take

[alamb]

I also think about solution 2 when handling that PR, and IMO solution 2 is the better option.

I agree solution 2 is better -- it just will take longer as it needs two coordinated PRs. One thing we have done in the past is do the initial implementation in DataFusion, and then file a ticket / port the code upstream to arrow-rs. Once the code is release in arrow-rs and available in Datafusion we remove the copy in DataFusion

[alamb]

I think this ticket is done now (though there is some follow up work to improve performance / port the code upstream)

Thanks again @tlm365