# ClickHouse release 26.10

DevFeed: [ClickHouse release 26.10](<https://devfeed.tech/articles/clickhouse-release-26-10-121768.md>)

Original publisher: [Read original article](<https://clickhouse.com/blog/clickhouse-release-26-10>)

Author: ClickHouse

Published: 2026-11-06T11:45:40Z

Content type: release

Language: en

Sources: [ClickHouse Blog](<https://devfeed.tech/sources/clickhouse-blog.md>)

Topics: [clickhouse](<https://devfeed.tech/topics/clickhouse.md>), [Release notes](<https://devfeed.tech/topics/release-notes.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [Functional programming](<https://devfeed.tech/topics/functional-programming.md>)

Tags: [cache](<https://devfeed.tech/tags/cache.md>), [clickhouse](<https://devfeed.tech/tags/clickhouse.md>), [composition](<https://devfeed.tech/tags/composition.md>), [experimental](<https://devfeed.tech/tags/experimental.md>), [faster](<https://devfeed.tech/tags/faster.md>), [flexible](<https://devfeed.tech/tags/flexible.md>), [freeform](<https://devfeed.tech/tags/freeform.md>), [function](<https://devfeed.tech/tags/function.md>), [more](<https://devfeed.tech/tags/more.md>), [release](<https://devfeed.tech/tags/release.md>)

## AI overview

ClickHouse 26.10 adds SQL composition with `|`, quota limits on inserted rows, and new array and map functions. Integer division can optionally truncate toward zero through `sql_compatible_division_for_integers`; that setting is off by default, errors on division by zero, and does not change stored views. Large `IN` subquery sets can spill to disk, enabling the example query to run under a 512 MiB limit with a 128 MiB spill threshold and 203 MiB peak use. Benchmarks report faster grouping by multiple `LowCardinality` columns and about 94% fewer rows read for Top-K grouped queries with dynamic filtering. The experimental columns cache reuses decoded column data (the example's repeated query falls from 32-34 ms to 12 ms); it currently supports wide MergeTree parts. Experimental `LATERAL JOIN` requires enabling a setting. Arbitrary `JOIN` conditions support patterns such as `LIKE`, but may use a block nested loop as slow as a cross join. The release also adds Freeform text input, SQLite input/output formats, and string filtering during scans; the example reports similar time but peak memory falling about 98%.

## Source excerpt

ClickHouse 26.10 introduces function composition, more flexible joins, faster GROUP BY queries, disk spilling for IN subqueries, an experimental columns cache, and Freeform and SQLite formats.