Skip to content

[python] Cache Blob indexes to avoid repeated footer reads - #9129

Draft
XiaoHongbo-Hope wants to merge 2 commits into
apache:masterfrom
XiaoHongbo-Hope:codex/blob-index-metadata-cache
Draft

[python] Cache Blob indexes to avoid repeated footer reads#9129
XiaoHongbo-Hope wants to merge 2 commits into
apache:masterfrom
XiaoHongbo-Hope:codex/blob-index-metadata-cache

Conversation

@XiaoHongbo-Hope

@XiaoHongbo-Hope XiaoHongbo-Hope commented Aug 9, 2026

Copy link
Copy Markdown
Contributor

Purpose

Follow up on #9125 by caching decoded Blob file indexes in the existing process-local file-format metadata cache.

When a worker reopens the same Blob file, it can reuse the cached lengths and offsets instead of reading and decoding the footer again. The cache remains bounded by file-format.metadata-cache.max-size; setting it to 0 b disables the cache.

Tests

  • blob_test.py
  • blob_table_test.py
  • parquet_metadata_cache_test.py
  • Python flake8

@XiaoHongbo-Hope
XiaoHongbo-Hope marked this pull request as ready for review August 9, 2026 09:05
@XiaoHongbo-Hope
XiaoHongbo-Hope marked this pull request as draft August 9, 2026 13:07
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant