Fix/parquet opener page index policy #19890
Open
+61
−4
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
github-actions
bot
added
core
Author
|
Resolved the issues! @kumarUjjawal |
aviralgarg05
force-pushed
the
fix/parquet-opener-page-index-policy
branch
from
91e0832 to
faffff0
Compare
martin-g
reviewed
Jan 20, 2026
| reader_metadata = load_page_index( | ||
| reader_metadata, | ||
| &mut async_file_reader, | ||
| // Since we're manually loading the page index the option here should not matter but we pass it in for consistency |
Member
There was a problem hiding this comment.
This comment probably needs to be updated.
| // Write parquet WITHOUT page index | ||
| let props = WriterProperties::builder().build(); | ||
|
|
||
| let file_fs = std::fs::File::create(&path).unwrap(); |
Member
There was a problem hiding this comment.
This tries to reopen the already opened file (line 979). Maybe you can just reuse it ?!
| let path = file.path().to_str().unwrap().to_string(); | ||
|
|
||
| // Write parquet WITHOUT page index | ||
| let props = WriterProperties::builder().build(); |
Member
There was a problem hiding this comment.
Which property exactly controls whether the page index is enabled or not ?
Maybe you can disable it explicitly here with WriterProperties::builder().with_xyz(false).build();, so it will still work as desired if it ever becomes enabled by default.
Member
There was a problem hiding this comment.
Is the update of the parquet-testing submodule needed for the new test ?
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Outdated suggestions cannot be applied.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.
Which issue does this PR close?
ParquetOpenerfails on files withoutPageIndexmetadata #19839.Rationale for this change
The ParquetOpener was using
ArrowReaderOptions::with_page_index(true), which internally setsPageIndexPolicy::Required. This caused sparse column chunk reads with row selection masks to fail with errors like "Invalid offset in sparse column chunk data" when reading Parquet files that lack page index metadata.Relaxing this policy to
PageIndexPolicy::Optionalallows DataFusion to gracefully handle files both with and without page index metadata while still leveraging the index when it exists.What changes are included in this PR?
PageIndexPolicy::Optionalinstead ofRequired.Are these changes tested?
Yes. I have added a dedicated regression test case:
This test writes a Parquet file specifically without page index metadata and verifies that ParquetOpener can read it successfully when
parquet_page_index_pruningis enabled.Are there any user-facing changes?
No. This is a bug fix that improves the robustness of the Parquet reader.