Improve performance of `regexp_count`

[alamb]

Is your feature request related to a problem or challenge?

@Omega359 and @xinlifoobar added the regexp_count function in https://github.com/apache/datafusion/pull/12970

However, regexp_count seems to be significantly slower than regexp_match and regexp_like. (see https://github.com/apache/datafusion/pull/12080#issuecomment-2316755718 for a benchmark)

Describe the solution you'd like

I would like to review the benchmark for regexp_count and figure out why it is slower and if so try to improve it

Describe alternatives you've considered

No response

Additional context

No response