Refactor regexplike signature

[jiashenC]

Which issue does this PR close?

Close some tasks in #13301.

What changes are included in this PR?

Simplify function signature.

Are these changes tested?

Added more tests for implicit casting.

Are there any user-facing changes?

No

[jiashenC]

@jayzhan211, I made some progress first for the RegExpLike function. In the original issue, you mentioned the signature can be replaced with Signature::string(2, Volatility::Immutable), but I do see this function can also accept 3 arguments. And this will fail tests like dataframe::dataframe_functions::test_fn_regexp_like. I wonder if I misunderstood something here?

[jayzhan211]

@jayzhan211, I made some progress first for the RegExpLike function. In the original issue, you mentioned the signature can be replaced with Signature::string(2, Volatility::Immutable), but I do see this function can also accept 3 arguments. And this will fail tests like dataframe::dataframe_functions::test_fn_regexp_like. I wonder if I misunderstood something here?

You can use one_of

signature: Signature::one_of(
                vec![
                    TypeSignature::String(2),
                    TypeSignature::String(3)
                ],
                Volatility::Immutable,
            ),

[jiashenC]

@jayzhan211 thanks for the reply! I mark this as ready for review.

[jayzhan211]

@jayzhan211 thanks for the reply! I mark this as ready for review.

You can run cargo test --test sqllogictests and ./dev/rust_lint.sh to cleanup CI

[alamb]

There appears to be a CI test failure in the examples were some queries no longer work

[alamb]

Thank you for this contribution @jiashenC

[Omega359]

I think the CI failures are a type casting issue. There are two rows in the example that have 4000 as the values that I think are not being coerced to strings as they should be.

 Error: Plan("Internal error: Failed to match any signature, errors: Error during planning: The signature expected 3 arguments but received 2,Error during planning: The signature expected NativeType::String but received NativeType::Int64.\nThis was likely caused by a bug in DataFusion's code and we would welcome that you file an bug report in our issue tracker No function matches the given name and argument types 'regexp_like(Utf8, Int64, Utf8)'. You might need to add explicit type casts.\n\tCandidate functions:\n\tregexp_like(String(2))\n\tregexp_like(String(3))")

Input data for the example: https://github.com/apache/datafusion/blob/main/datafusion/physical-expr/tests/data/regex.csv

[alamb]

Marking as draft as I think this PR is no longer waiting on feedback. Please mark it as ready for review when it is ready for another look

I am trying to make the review backlog easier to understand

[alamb]

(sorry I accidnetally closed the PR)

[jayzhan211]

Thanks @jiashenC @alamb @Omega359