Skip to content

[C++] UpdateLevelHistogram produces incorrect histograms in fast path #50778

Closed

Description

@wecharyu

Describe the bug, including details regarding any error messages, version, and platform.

parquet::UpdateLevelHistogram produces incorrect histograms when max_level == 1 and the input contains more than 16,384 levels.

The fast path processes input in chunks, but each iteration accumulates from levels.begin() instead of the current chunk iterator. As a result, the first chunk is counted repeatedly and later chunks are skipped.

Component(s)

C++

Activity

  1. added a commit that references this issue on Aug 31, 2026
  2. added this to the 26.0.0 milestone on Aug 31, 2026
  3. pitrou commented on Aug 31, 2026

    @pitrou
    Member

    Issue resolved by pull request 50780
    #50780

  4. pitrou commented on Aug 31, 2026

    @pitrou
    Member

    Thanks for diagnosing and fixing this embarrassing issue @wecharyu !

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

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions