Skip to content

[BUG] Parquet chunked reader can throw an 'unexpected short subpass' exception under certain conditions. #18043

Open
@nvdbaranec

Description

This was discovered as a byproduct of changes to nvcomp temporary memory usage for decompression. The change caused us to produce a slightly different set of chunks, exposing the underlying bug in the chunked reader itself (nvcomp was not doing anything wrong). Spark Rapids customers have experienced this as well, under difficult-to-reproduce conditions, so having a clean repro case here is nice.

To reproduce, build cudf using nvcomp 4.2.0.11 (#18042) and run the tests. Two of the list tests, ParquetChunkedReaderInputLimitConstrainedTest.MixedColumns and ParquetChunkedReaderInputLimitTest.List will throw the exception.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

Labels

SparkFunctionality that helps Spark RAPIDSbugSomething isn't working

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions