[SPARK-48019] Fix incorrect behavior in ColumnVector/ColumnarArray with dictionary and nulls #46254
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
|
thanks, merging to master/3.5! |
cloud-fan
pushed a commit
that referenced
this pull request
Apr 28, 2024
…th dictionary and nulls This fixes how `ColumnVector` handles copying arrays when the vector has a dictionary and null values. The possible issues with the previous implementation: - An `ArrayIndexOutOfBoundsException` may be thrown when the `ColumnVector` has nulls and dictionaries. This is because the dictionary id for `null` entries might be invalid and should not be used for `null` entries. - Copying a `ColumnarArray` (which contains a `ColumnVector`) is incorrect, if it contains `null` entries. This is because copying a primitive array does not take into account the `null` entries, so all the null entries get lost. These changes are needed to avoid `ArrayIndexOutOfBoundsException` and to produce correct results when copying `ColumnarArray`. The only user facing changes are to fix existing errors and incorrect results. Added new unit tests. No. Closes #46254 from gene-db/dictionary-nulls. Authored-by: Gene Pang <[email protected]> Signed-off-by: Wenchen Fan <[email protected]> (cherry picked from commit 76ce6b0) Signed-off-by: Wenchen Fan <[email protected]>
cloud-fan
pushed a commit
that referenced
this pull request
May 5, 2024
…when nulls exist ### What changes were proposed in this pull request? This is a followup to #46254 . Instead of using object arrays when nulls are present, continue to use primitive arrays when appropriate. This PR sets the null bits appropriately for the primitive array copy. Primitive arrays are faster than object arrays and won't create unnecessary objects. ### Why are the changes needed? This will improve performance and memory usage, when nulls are present in the `ColumnarArray`. ### Does this PR introduce _any_ user-facing change? This is expected to be faster when copying `ColumnarArray`. ### How was this patch tested? Existing tests. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #46372 from gene-db/primitive-nulls. Authored-by: Gene Pang <[email protected]> Signed-off-by: Wenchen Fan <[email protected]>
cloud-fan
pushed a commit
that referenced
this pull request
May 5, 2024
…when nulls exist ### What changes were proposed in this pull request? This is a followup to #46254 . Instead of using object arrays when nulls are present, continue to use primitive arrays when appropriate. This PR sets the null bits appropriately for the primitive array copy. Primitive arrays are faster than object arrays and won't create unnecessary objects. ### Why are the changes needed? This will improve performance and memory usage, when nulls are present in the `ColumnarArray`. ### Does this PR introduce _any_ user-facing change? This is expected to be faster when copying `ColumnarArray`. ### How was this patch tested? Existing tests. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #46372 from gene-db/primitive-nulls. Authored-by: Gene Pang <[email protected]> Signed-off-by: Wenchen Fan <[email protected]> (cherry picked from commit bf2e254) Signed-off-by: Wenchen Fan <[email protected]>
Member
There was a problem hiding this comment.
Thank you, @gene-db and @cloud-fan .
SPARK-48019 was filed with Affected Version: 3.4.3. I also verified that the test cases fail on branch-3.4. Since this is a correctness issue, let me backport this and the follow-up to branch-3.4
dongjoon-hyun
pushed a commit
that referenced
this pull request
Sep 4, 2024
…th dictionary and nulls This fixes how `ColumnVector` handles copying arrays when the vector has a dictionary and null values. The possible issues with the previous implementation: - An `ArrayIndexOutOfBoundsException` may be thrown when the `ColumnVector` has nulls and dictionaries. This is because the dictionary id for `null` entries might be invalid and should not be used for `null` entries. - Copying a `ColumnarArray` (which contains a `ColumnVector`) is incorrect, if it contains `null` entries. This is because copying a primitive array does not take into account the `null` entries, so all the null entries get lost. These changes are needed to avoid `ArrayIndexOutOfBoundsException` and to produce correct results when copying `ColumnarArray`. The only user facing changes are to fix existing errors and incorrect results. Added new unit tests. No. Closes #46254 from gene-db/dictionary-nulls. Authored-by: Gene Pang <[email protected]> Signed-off-by: Wenchen Fan <[email protected]> (cherry picked from commit 76ce6b0) Signed-off-by: Wenchen Fan <[email protected]>
dongjoon-hyun
pushed a commit
that referenced
this pull request
Sep 4, 2024
…when nulls exist ### What changes were proposed in this pull request? This is a followup to #46254 . Instead of using object arrays when nulls are present, continue to use primitive arrays when appropriate. This PR sets the null bits appropriately for the primitive array copy. Primitive arrays are faster than object arrays and won't create unnecessary objects. ### Why are the changes needed? This will improve performance and memory usage, when nulls are present in the `ColumnarArray`. ### Does this PR introduce _any_ user-facing change? This is expected to be faster when copying `ColumnarArray`. ### How was this patch tested? Existing tests. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #46372 from gene-db/primitive-nulls. Authored-by: Gene Pang <[email protected]> Signed-off-by: Wenchen Fan <[email protected]> (cherry picked from commit bf2e254) Signed-off-by: Wenchen Fan <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
This fixes how
ColumnVectorhandles copying arrays when the vector has a dictionary and null values. The possible issues with the previous implementation:ArrayIndexOutOfBoundsExceptionmay be thrown when theColumnVectorhas nulls and dictionaries. This is because the dictionary id fornullentries might be invalid and should not be used fornullentries.ColumnarArray(which contains aColumnVector) is incorrect, if it containsnullentries. This is because copying a primitive array does not take into account thenullentries, so all the null entries get lost.Why are the changes needed?
These changes are needed to avoid
ArrayIndexOutOfBoundsExceptionand to produce correct results when copyingColumnarArray.Does this PR introduce any user-facing change?
The only user facing changes are to fix existing errors and incorrect results.
How was this patch tested?
Added new unit tests.
Was this patch authored or co-authored using generative AI tooling?
No.