sarutak opened a new pull request, #57213:
URL: https://github.com/apache/spark/pull/57213
### What changes were proposed in this pull request?
This PR backports SPARK-57932 (#56998) for `branch-4.2` with conflict
resolution.
`regexp_instr` returned matcher.start() + 1 (a UTF-16 code-unit index), and
`regexp_replace`'s position argument was used as a UTF-16 offset in
matcher.region(pos-1, ...) and its out-of-range guard. Both now use code-point
positions via codePointCount/offsetByCodePoints.
### Why are the changes needed?
Spark string positions are 1-based code points (instr, locate, position,
substring, length). For strings containing supplementary characters (code
points > U+FFFF, e.g. emojis), these two functions gave wrong results:
- regexp_instr('😀ab', 'ab') returned 3 instead of 2.
- regexp_replace('😀aXa', 'a', 'Z', 3) returned 😀ZXZ instead of 😀aXZ
(replaced a match before position); regexp_replace('😀😀', '😀', 'Z', 3) returned
😀Z instead of the input unchanged. This is the sibling of the LIKE fix in
SPARK-55453.
### Does this PR introduce _any_ user-facing change?
Yes. regexp_instr and regexp_replace now return code-point-consistent
results for strings with supplementary characters.
### How was this patch tested?
Added UT.
### Was this patch authored or co-authored using generative AI tooling?
Yes.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
