[C++] SQL-style glob string match kernel

[asfimport]

Similar to the SQL LIKE and ILIKE operators, with wildcard _ matching a single character and wildcard % matching multiple characters, and an option controlling case sensitivity

Reporter: Ian Cook / @ianmcook
Assignee: David Li / @lidavidm

Related issues:

• [C++] String algorithm library for StringArray/BinaryArray (is a child of)
• [C++] Add string starts-with/ends-with kernels (is related to)
• [R] Bindings for stringr::str_like (is depended upon by)

PRs and other links:

• GitHub Pull Request #10356

_{Note: This issue was originally created as ARROW-12715. Please see the migration documentation for further details.}

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
This is kind of a "simpler" version compared to using an actual regex expression?

I am wondering if we need to exactly implement this as a separate kernel with those specific features, or if this could be written in terms of the match_substring_regex. In principle an input string to match could be transformed into the equivalent regex pattern? And then we don't need to implement a light form of pattern matching ourselves.

(of course, that would then require arrow to be compiled with a regex library for this function, and a specialized version might be faster?)

[asfimport]

Ian Cook / @ianmcook:
@jorisvandenbossche  yes; I would suggest looking at the implementation of LIKE in Impala to see what they did there. Like Arrow, Impala is implemented in C++ and uses RE2, and it also has regex match functions and operators.

[asfimport]

David Li / @lidavidm:
Impala checks if the pattern can be converted into a prefix/suffix/substring/exact match search, or else converts the pattern into a regex: https://github.com/apache/impala/blob/9c38568657d62b6f6d7b10aa1c721ba843374dd8/be/src/exprs/like-predicate.cc#L63-L108

We could do something similar.

[asfimport]

Ian Cook / @ianmcook:
Options for case insensitivity will be a follow-up (ARROW-12835)

[asfimport]

David Li / @lidavidm:
Issue resolved by pull request 10356
#10356