clickhouse
https://github.com/yandex/clickhouse
C++
ClickHouse is a free analytic DBMS for big data.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
C++ not yet supported68 Subscribers
Add a CodeTriage badge to clickhouse
Help out
- Issues
- RFC: Lifetime-segregated memory placement (attribution == placement)
- JSON object traversal pays duplicate-key bookkeeping on every key of every object — a no-duplicate fast path removes it (json_type_parsing q3 +33%, 26.6 → 26.7, aarch64)
- JSON parsing: make `JSONExtractTree` object traversal allocation-free per key (reusable path buffer + two-bit duplicate-key masks) — JSONAsObject workloads parse 25–34% faster, peak memory down (pre-existing cost, not a regression)
- Keyed aggregations over huge constant-folded literals still pay a per-execution O(N) content-hash in the hash-table-stats cache key (bounded residual of #107643; ~9ms per execution for a 5M-element literal)
- Tolerate a concurrently dropped namespace when listing an Iceberg REST catalog
- Fix PromQL sgn() for NaN and negative zero
- Correlated-subquery decorrelation with in-memory buffer disabled: `Cannot clone ReadFromLocalReplica plan step` under parallel replicas
- Add JOIN observability columns to system.query_log
- Run the FileLog directory watch on its own thread
- Custom-key parallel replicas over a single-shard cluster set `distributed_group_by_no_merge = 2`, so the initiator never merges the per-replica aggregates
- Docs
- C++ not yet supported