fix: Add memory management for approx_most_frequent addSingleGroupRawInput #15852
+130
−18
Conversation
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
✅ Deploy Preview for meta-velox canceled.
|
meta-cla
bot
added
the
CLA Signed
|
@duxiao1212 has exported this pull request. If you are a Meta employee, you can view the originating Diff in D89728171. |
…Input (facebookincubator#15852) Summary: Add memory management to approx_most_frequent aggregate function to prevent OOM during global aggregation, where spilling is not effective. Changes: 1. Modified ApproxMostFrequentStreamSummary::insert() to return eviction info (evicted value and count) 2. Added memory tracking (liveBytes_, deadBytes_) for StringView accumulator 3. Added rebuild() to compact string storage when dead bytes exceed threshold 4. Added per-key size limit to prevent individual oversized keys Motivation: For queries like: SELECT APPROX_MOST_FREQUENT(20, CAST(feature AS JSON), 1000), ... FROM table LIMIT 10000000 With 10M input rows and ~82KB per row, without memory management: **20251215_232949_18211_b5kzi** The HashStringAllocator accumulates string storage for all inserted keys When keys are evicted from the Space-Saving summary, their string storage becomes "dead" but is never freed Memory grows unbounded until OOM This fix addresses: Single large keys - Per-key size limit (kMaxTotalStringBytes / capacity) prevents any individual key from consuming excessive memory Dead memory accumulation - rebuild() triggers when deadBytes > kMaxTotalStringBytes * 0.25, compacting storage by copying only live keys to new allocation and freeing the old storage Differential Revision: D89728171
duxiao1212
force-pushed
the
export-D89728171
branch
from
807ffad to
3a0bb48
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Summary:
Add memory management to approx_most_frequent aggregate function to prevent OOM during global aggregation, where spilling is not effective.
Changes:
Motivation:
For queries like:
SELECT APPROX_MOST_FREQUENT(20, CAST(feature AS JSON), 1000), ...
FROM table LIMIT 10000000
With 10M input rows and ~82KB per row, without memory management:
20251215_232949_18211_b5kzi
The HashStringAllocator accumulates string storage for all inserted keys
When keys are evicted from the Space-Saving summary, their string storage becomes "dead" but is never freed
Memory grows unbounded until OOM
This fix addresses:
Single large keys - Per-key size limit (kMaxTotalStringBytes / capacity) prevents any individual key from consuming excessive memory
Dead memory accumulation - rebuild() triggers when deadBytes > kMaxTotalStringBytes * 0.25, compacting storage by copying only live keys to new allocation and freeing the old storage
Differential Revision: D89728171