Conversation
geoffreyclaude
force-pushed
the
geoffrey.claude/concurrent_heads
branch
from
c0290b4 to
6da3517
Compare
When scanning an exact list of remote Parquet files, the ListingTable was fetching file metadata (via head calls) sequentially. This was due to using `stream::iter(file_list).flatten()`, which processes each one-item stream in order. For remote blob stores, where each head call can take tens to hundreds of milliseconds, this sequential behavior significantly increased the time to create the physical plan. This commit replaces the sequential flattening with concurrent merging using `tream::iter(file_list).flatten_unordered(meta_fetch_concurrency). With this change, the `head` requests are executed in parallel (up to the configured `meta_fetch_concurrency` limit), reducing latency when creating the physical plan. Note that the ordering loss introduced by `flatten_unordered` is perfectly acceptable as the file list will anyways be fully sorted by path in `split_files` before being returned. Additionally, tests have been updated to ensure that metadata fetching occurs concurrently.
geoffreyclaude
force-pushed
the
geoffrey.claude/concurrent_heads
branch
from
6da3517 to
bbfd5b0
Compare
gabotechs
approved these changes
Mar 3, 2025
gabotechs
pushed a commit
that referenced
this pull request
Mar 4, 2025
…#2) When scanning an exact list of remote Parquet files, the ListingTable was fetching file metadata (via head calls) sequentially. This was due to using `stream::iter(file_list).flatten()`, which processes each one-item stream in order. For remote blob stores, where each head call can take tens to hundreds of milliseconds, this sequential behavior significantly increased the time to create the physical plan. This commit replaces the sequential flattening with concurrent merging using `tream::iter(file_list).flatten_unordered(meta_fetch_concurrency). With this change, the `head` requests are executed in parallel (up to the configured `meta_fetch_concurrency` limit), reducing latency when creating the physical plan. Note that the ordering loss introduced by `flatten_unordered` is perfectly acceptable as the file list will anyways be fully sorted by path in `split_files` before being returned. Additionally, tests have been updated to ensure that metadata fetching occurs concurrently.
gabotechs
pushed a commit
that referenced
this pull request
Mar 4, 2025
…#2) When scanning an exact list of remote Parquet files, the ListingTable was fetching file metadata (via head calls) sequentially. This was due to using `stream::iter(file_list).flatten()`, which processes each one-item stream in order. For remote blob stores, where each head call can take tens to hundreds of milliseconds, this sequential behavior significantly increased the time to create the physical plan. This commit replaces the sequential flattening with concurrent merging using `tream::iter(file_list).flatten_unordered(meta_fetch_concurrency). With this change, the `head` requests are executed in parallel (up to the configured `meta_fetch_concurrency` limit), reducing latency when creating the physical plan. Note that the ordering loss introduced by `flatten_unordered` is perfectly acceptable as the file list will anyways be fully sorted by path in `split_files` before being returned. Additionally, tests have been updated to ensure that metadata fetching occurs concurrently.
gabotechs
changed the title
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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?
Rationale for this change
When scanning an exact list of remote Parquet files, the ListingTable was fetching file metadata (via head calls) sequentially. This was due to using
stream::iter(file_list).flatten(), which processes each one-item stream in order. For remote blob stores, where each head call can take tens to hundreds of milliseconds, this sequential behavior significantly increased the time to create the physical plan.What changes are included in this PR?
This commit replaces the sequential flattening with concurrent merging using
stream::iter(file_list).flatten_unordered(meta_fetch_concurrency). With this change, theheadrequests are executed in parallel (up to the configuredmeta_fetch_concurrencylimit), reducing latency when creating the physical plan.Note that the ordering loss introduced by
flatten_unorderedis perfectly acceptable as the file list will anyways be fully sorted by path insplit_filesbefore being returned.Are these changes tested?
No tests for this fork to mitigate future git conflict risks. Tests have been updated in the upstream PR apache#14918.
Are there any user-facing changes?
No user-facing changes besides reducing the latency in this particular situation.