Add nilike support in comparison
#1846
Conversation
Codecov Report
@@ Coverage Diff @@
## master #1846 +/- ##
==========================================
+ Coverage 83.49% 83.50% +0.01%
==========================================
Files 201 201
Lines 56903 56951 +48
==========================================
+ Hits 47511 47559 +48
Misses 9392 9392
Continue to review full report at Codecov.
|
MazterQyou
force-pushed
the
upstream-patch/comparison-nilike
branch
from
49b5eae
to
45b524e
Compare
| result.append( | ||
| !left | ||
| .value(i) | ||
| .to_uppercase() |
There was a problem hiding this comment.
Given this allocates a new string, I wonder how much faster this fast path actually is? Do you have some benchmark results?
There was a problem hiding this comment.
I have simply re-used this code from the already existing ilike comparator function, so I have no idea regarding that. I agree, it does make sense to do some benchmarks. Although I feel ends_with and starts_with sound simple enough to be much faster than a regexp match implicitly.
| @@ -3984,6 +4067,60 @@ mod tests { | |||
| vec![false, true, false, false] | |||
| ); | |||
|
|
|||
| test_utf8!( | |||
There was a problem hiding this comment.
Maybe a test of non-ASCII characters, e.g. Â â or something
| /// [`LargeStringArray`] and a scalar. | ||
| /// | ||
| /// See the documentation on [`like_utf8`] for more details. | ||
| pub fn nilike_utf8_scalar<OffsetSize: OffsetSizeTrait>( |
There was a problem hiding this comment.
It occurs to me that we now have four methods ilike_utf8_scalar, nlike_utf8_scalar, like_utf8_scalar, nilike_utf8_scalar where the only difference appears to be an optional post-conversion on left.value(i). Perhaps we could refactor out the common logic into a function taking F: Fn(&str) -> Cow<'_, str> or something?
There was a problem hiding this comment.
I had this in mind as well but decided to open the PR nevertheless considering there are already three similar functions (which means whoever accepted them in was okay with that). I'll refactor those and see how that goes.
|
Would you like me to get this PR in and we can maybe follow up with feedback in another PR? I think as is this is a good step forward? Edit: merging master should fix the MIRI failures |
|
I am up for making the changes you suggested, although if you think it's good to have this faster and improve on later, I can rebase now. |
|
Merging this one in to keep the queue down -- Thank you @MazterQyou and @tustvold . I would love to see a refactor PR that reduced the code duplication |
Which issue does this PR close?
Closes #1845.
Rationale for this change
comparisonkernel has functionslikeandnlike, which performLIKEandNOT LIKEoperations. There was also an addition ofilikeback in late 2021, which allows to performeILIKE(case-insensitiveLIKE) operations. However,nilike, which is whatnliketolikeis, respectively (NOT ILIKE), has been missing. It is a useful addition to perform similarNOToperation with already supportedILIKE.What changes are included in this PR?
This PR adds
nilikefunctions tocomparisonkernel that perform SQLleft NOT ILIKE rightoperation on arrays and scalars, as well as related tests and benches.Are there any user-facing changes?