ClickHouse roadmap 2026 (draft)

[alexey-milovidov]

This is the ClickHouse roadmap 2026.

This roadmap does not cover the tasks related to infrastructure, orchestration, documentation, marketing, external integrations, drivers, etc.

See also:

Roadmap 2025: #74046
Roadmap 2024: #58392
Roadmap 2023: #44767
Roadmap 2022: #32513
Roadmap 2021: #17623
Roadmap 2020: link
Roadmap 2019: link
Roadmap 2018: link

Query Engine

• Parallel replicas for Merge tables Parallel replicas are not implemented for Merge tables #67770
• Materialized CTE Materialized CTE #61086
• Optimization by combining aggregation with topK
• Optimization of aggregation with flat combining Speed up GROUP BY for large states #79428
• Cache for deserialized columns
• Cascades cost-based optimizer for JOINs
• Using partition key to optimize JOINs
• Using partition key to optimize DISTINCT, IN, GROUP BY, and LIMIT BY
• Runtime filters for JOINs by default
• Optimization of anti-joins: LEFT JOIN ... WHERE ... IS NULL to NOT IN
• Deriving index condition from the right-hand side of INNER JOIN
• Support for certain SQL functions without parentheses Allow certain functions without parentheses in SQL #52102
• Optimizations for a single dictionary in LowCardinality [RFC] Use global LowCardinality dictionary for optimizations if it is small enough #72717
• Block-level hints WIP Add Sorting Optimization with General Data Hints #48800
• Query cache for subqueries Can the features of Query Cache support subqueries? #51656
• Query cache for partial results Use query cache when a query shares most of the query pipeline. #57490
• On-disk query cache Allow using cache disk for query result cache. #52141
• Improvements for the query condition cache
• Streaming queries Streaming queries model with cursors #63312
• Enable JIT compilation on macOS

Resiliency

• Automatic external memory for DISTINCT and LIMIT BY
• Automatic external memory for IN
• Automatic external memory for JOIN
• Memory configuration in the workload scheduler [WIP] Workload scheduling: Memory reservations #82414
• GROUP BY in order by default
• Automatic selection of max_insert_threads
• Tracking the space in external memory
• Spill to S3 in addition to spill to the cache
• Allow the cancellation of scalar subqueries

Data Storage

• Text index in GA
• Unique key constraint RFC: Unique key constraint #70589
• Automatic LowCardinality columns Automatic LowCardinality representation of columns #69916
• Compact statistics storage
• Statistics by default
• Virtual shards on top of shared storage ☁
• ShardedMap data type Add shards to data type Map #47045
• Transactions for Replicated tables
• Scoring in the text index

Data Lakes

• Metadata caching for Parquet files
• Support arbitrary nesting in database and table names Support arbitrary nesting in database and table names. #71171
• Support for case-insensitive database, table, and column names A setting to allow case insensitive column, table and database names. #33935
• Support for deletion vectors in Iceberg v3
• Query Condition Cache for data lakes
• Lazy columns reading for data lakes
• Data Lake catalogs in GA
• Automatic cluster functions for subqueries
• Support for AWS S3 Tables
• Support for Google BigLake
• Plain Rewritable metadata storage to support hard links
• Materialized views on top of data lakes
• Background merges for data lakes ☁

Interfaces

• Allow interactive queries to finish in the background Allow interactive queries finish in the background #49683
• Configurable query rewrite rules Query rewrite rules (RFC) #80084
• PromQL dialect in GA
• Support for Arrow Flight SQL
• Kafka2 in GA
• Configuration of query handlers with SQL
• Remote and Cloud database engines Remote database engine #59304
• Drivers for UDFs Drivers for User-Defined Functions #71172
• WebAssembly UDFs
• API for query construction Add combining http parameters with queries, new handler for accessing tables as files  #64336
• Even simpler data upload Even simpler data upload. #38775
• Passing input data to the url table function Support input data for reading from the URL table engine #45994
• Framing formats for query results
• Predictive autocomplete in the CLI Autocomplete #69641

Cleanups

• Remove old analyzer
• Remove old predicate pushdown mechanics
• Remove the hypotheses index
• Remove Parquet reader v1
• Remove or fix Kusto (KQL)
• Remove Window View in favor of streaming queries

[CurtizJ]

@alexey-milovidov

Can we include improvements of the query condition cache in the roadmap (#89169 and #78502)?

[alexey-milovidov]

@CurtizJ, sure, added. I'm a bit afraid because both of these tasks are underspecified - let's create a good RFC.