ClickHouse Release 26.1
6 hours ago
- #database
- #ClickHouse
- #optimization
- ClickHouse 26.1 introduces 25 new features, 43 performance optimizations, and 176 bug fixes.
- New system table `system.zookeeper_info` provides detailed status of all Keeper clusters.
- Function `reverseBySeparator` allows reversing delimited collections without converting to an array.
- New function `mergeTreeAnalyzeIndexes` aids in distributed index analysis with parallel replicas.
- Projections can now be defined with an INDEX clause for better query performance.
- Support for Nullable Tuple types enhances compatibility with external data formats.
- New data type `QBit` for vector embeddings with runtime precision tuning.
- Optimizations for GROUP BY and DISTINCT operations over LowCardinality columns.
- JIT compilation extended to more functions, including aggregates over Nullable, Decimal, and big integers.
- Memory reduction for inserts into wide tables, demonstrated with `system.metric_log`.
- New `concurrent_threads_scheduler` configuration improves latency distribution for highly concurrent queries.
- `sparseGrams` tokenizer enhances text search efficiency by using longer, more specific ngrams.
- Keeper API now supports HTTP control for better management.
- Asynchronous inserts now support exactly-once semantics for materialized views.
- ClickHouse Kubernetes operator automates cluster provisioning, scaling, and upgrades.
- Upcoming events include meetups in Brussels, Barcelona, London, and more.