Implement native support StringView for overlay #11968
Conversation
Signed-off-by: Chojan Shang <psiace@apache.org>
Signed-off-by: Chojan Shang <psiace@apache.org>
Signed-off-by: Chojan Shang <psiace@apache.org>
| 01)Projection: overlay(CAST(test.column1_utf8view AS Utf8), Utf8("foo"), Int64(2)) AS c1 | ||
| 01)Projection: overlay(test.column1_utf8view, Utf8View("foo"), Int64(2)) AS c1 | ||
| 02)--TableScan: test projection=[column1_utf8view] | ||
|
|
There was a problem hiding this comment.
Likewise, can we please also add actually running these queries to the tests
like
query
SELECT OVERLAY(column1_utf8view PLACING 'foo' FROM 2 ) as c1?
Signed-off-by: Chojan Shang <psiace@apache.org>
| } | ||
| } | ||
|
|
||
| macro_rules! process_overlay { |
There was a problem hiding this comment.
It would be nice to have some sort of trait in arrow-rs that allowed us to write this as a generic function
It actually does have
https://github.com/apache/arrow-rs/blob/2461a16c19ee5032531b1c05dd7e7192bc842e0f/arrow-string/src/like.rs#L158-L161
But that is not public
@XiangpengHao do you know of anything that is pub?
We could also implement such a trait for DataFusion's convenience, and then propose upstreaming it 🤔
There was a problem hiding this comment.
I was just behind -- it seems that @Omega359 did exactly this in StringArrayType #11941
datafusion/datafusion/functions/src/unicode/lpad.rs
Lines 251 to 263 in 94034be
Maybe we can start to pull that trait into its own module and start reusing it across the string functions 🤔
Also, there is the ArrayAccessor pattern used elegantly by @devanbenz in #11967 🤔
There was a problem hiding this comment.
The current style is inspired by rpad in #11942 . I'll be rewriting it with ArrayAccessor, which I've used in other PRs, and it's a much more elegant way.
|
Thanks @PsiACE |
* Implement native support StringView for overlay Signed-off-by: Chojan Shang <psiace@apache.org> * Re-write impl of overlay Signed-off-by: Chojan Shang <psiace@apache.org> * Minor update Signed-off-by: Chojan Shang <psiace@apache.org> * Add more tests Signed-off-by: Chojan Shang <psiace@apache.org> --------- Signed-off-by: Chojan Shang <psiace@apache.org>
Which issue does this PR close?
Closes #11909 .
Rationale for this change
Ref: #11942
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?