Skip to content

Error Reading Page Index When Not Available  #2434

@tustvold

Description

@tustvold

Describe the bug

The doc comment for read_pages_locations reads

If not the format not available return an empty vector.

I would take this to mean it would return an empty vector if index information is not available, instead it returns

General("underlying Thrift error: end of file")

To Reproduce

Enable ReadOptions::enable_page_index and try to read a parquet file that lacks a PageIndex

Expected behavior

Enabling the page index shouldn't require it to exist

Additional context

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugparquetChanges to the parquet crate

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions