Fix ilike_utf8_scalar kernals #2545
Conversation
| @@ -468,7 +468,7 @@ pub fn ilike_utf8_scalar<OffsetSize: OffsetSizeTrait>( | |||
| if !right.contains(is_like_pattern) { | |||
| // fast path, can use equals | |||
| for i in 0..left.len() { | |||
| result.append(left.value(i) == right); | |||
| result.append(left.value(i).to_uppercase() == right.to_uppercase()); | |||
There was a problem hiding this comment.
This needs to be done once only for the right scalar?
There was a problem hiding this comment.
Sorry, A bit confused.
Did you mean that the right.to_uppercase() and be extracted to outside the if block rather than doing it every time in the if else blocks ?
There was a problem hiding this comment.
Yes, and also outside the inner loop. The compiler often won't find out that it's possible to do this.
There was a problem hiding this comment.
Fixed in the latest commit
| @@ -551,7 +551,7 @@ pub fn nilike_utf8_scalar<OffsetSize: OffsetSizeTrait>( | |||
| if !right.contains(is_like_pattern) { | |||
| // fast path, can use equals | |||
| for i in 0..left.len() { | |||
| result.append(left.value(i) != right); | |||
| result.append(left.value(i).to_uppercase() != right.to_uppercase()); | |||
|
Benchmark runs are scheduled for baseline = 89f2e45 and contender = b2f0c65. b2f0c65 is a master commit associated with this PR. Results will be available as each benchmark for each run completes. |
Which issue does this PR close?
Closes #2544 .
Rationale for this change
Incorrect logic in equals path of ilike_utf8_scalar kernals
What changes are included in this PR?
Bug fix
Are there any user-facing changes?