perf: show the most frequent text index values, in batches of 10 keys by knudtty · Pull Request #3617 · ClickHouse/ClickStack

fix: drop the row cap from text index value lookups

knudtty

committed

326c149

feat: show the most frequent text index values with topKWeighted

knudtty

committed

cb5ed9a

perf: read text index values per key with a token cap

knudtty

committed

bddaf87

perf: run text index values in batches of 10 keys

knudtty

committed

54f2bf6

perf: read text index values in 1024-row blocks so per-key LIMITs stop early

knudtty

committed

a8806af

Revert "perf: read text index values in 1024-row blocks so per-key LIMITs stop early"

knudtty

committed

7e32a81

Revert "perf: read text index values per key with a token cap"

knudtty

committed

5fe1183

Reapply "perf: read text index values per key with a token cap"

knudtty

committed

dde2feb

docs: describe per-key text index value batches in the changeset

knudtty

committed

7ec3fe1

perf: use 4096-token dictionary blocks for key=value text indexes in seed schemas

knudtty

committed

4639a46

perf: read at most 20,000 text index tokens per key

knudtty

committed

363ee04

perf: use 8192-token dictionary blocks for every text index

knudtty

committed

892f472

perf: read Map text index values per key only when dictionary blocks are large

knudtty

committed

0634f1a

perf: rank small-block text index values over every token read

knudtty

committed

94ecf87

perf: keep main's batching for small-block Map text index values

knudtty

committed

45a166f

perf: keep main's query and row cap for small-block Map text index values

knudtty

committed

09028d6

perf: batch small-block Map text index values by 10 keys

knudtty

committed

8f0b114

docs: fold the row cap change into the text index values changeset

knudtty

committed

0d5c695

Navigation Menu

Repository navigation

perf: show the most frequent text index values, in batches of 10 keys - #3617

perf: show the most frequent text index values, in batches of 10 keys#3617
knudtty wants to merge 18 commits into
claude/map-key-index-median-partsfrom
claude/topk-text-index-values

Commits on Nov 13, 2026

fix: drop the row cap from text index value lookups

feat: show the most frequent text index values with topKWeighted

perf: read text index values per key with a token cap

perf: run text index values in batches of 10 keys

perf: read text index values in 1024-row blocks so per-key LIMITs stop early

Revert "perf: read text index values in 1024-row blocks so per-key LIMITs stop early"

Revert "perf: read text index values per key with a token cap"

Reapply "perf: read text index values per key with a token cap"

docs: describe per-key text index value batches in the changeset

perf: use 4096-token dictionary blocks for key=value text indexes in seed schemas

perf: read at most 20,000 text index tokens per key

perf: use 8192-token dictionary blocks for every text index

perf: read Map text index values per key only when dictionary blocks are large

perf: rank small-block text index values over every token read

perf: keep main's batching for small-block Map text index values

perf: keep main's query and row cap for small-block Map text index values

perf: batch small-block Map text index values by 10 keys

docs: fold the row cap change into the text index values changeset