analyze: add HLLSketch for per-region singleton NDV estimation#412
Closed
0xPoe wants to merge 2 commits into
Closed
analyze: add HLLSketch for per-region singleton NDV estimation#4120xPoe wants to merge 2 commits into
0xPoe wants to merge 2 commits into
Conversation
Add a HLLSketch message and hll_ndv_sketch/hll_singleton_sketch fields to RowSampleCollector. These let TiDB estimate the global singleton (f1) count via a per-region leave-one-out using fixed-size HyperLogLog sketches, instead of retaining large per-region FM sketches. The FM sketch path is unchanged for the global/stored NDV. Signed-off-by: 0xPoe <techregister@pm.me>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Adds a
HLLSketchmessage andhll_ndv_sketch/hll_singleton_sketchrepeated fields toRowSampleCollector.These let TiDB estimate the global singleton (f1) count for ANALYZE NDV sub-sampling via a per-region leave-one-out using fixed-size HyperLogLog sketches, instead of retaining large per-region FM sketches (which is O(regions) memory and spiked to ~20 GiB at ~6700 regions). The FM sketch path is unchanged and still backs the global/stored NDV.
Notes
proto: add NDV rate to analyze requestcommit (also in this branch).len(registers) == 1<<precision. Rust (TiKV) and Go (TiDB) use an identical register layout so a TiKV-built sketch reads correctly in TiDB.