{ "schema_version": 2, "id": "operate/oss_and_stack/stack-with-enterprise/release-notes/redisearch/redisearch-2.6-release-notes", "title": "RediSearch 2.6 release notes", "url": "https://redis.io/docs/latest/operate/oss_and_stack/stack-with-enterprise/release-notes/redisearch/redisearch-2.6-release-notes/", "summary": "Search using wildcard queries for TEXT and TAG fields, multi-value indexing and querying of attributes for any attribute type, and indexing double-precision floating-point vectors and range queries from a given vector.", "tags": [ "docs", "operate", "stack" ], "last_updated": "2026-09-21T16:36:33+01:00", "page_type": "content", "content_hash": "eb1ff3f45a25c4352b321a1bb59af71d9fc0432103ea9839667dec7fe8f8421f", "sections": [ { "id": "requirements", "title": "Requirements", "role": "overview", "text": "RediSearch v2.6.37 requires:\n\n- Minimum Redis compatibility version (database): 6.0.16\n- Minimum Redis Enterprise Software version (cluster): 6.2.8" }, { "id": "v2637-june-2026", "title": "v2.6.37 (June 2026)", "role": "content", "text": "This is a maintenance release for Redis Search 2.6.\nUpdate urgency: `LOW`: No need to upgrade unless there are new features you want to use.\n\n**Bug Fixes**\n\n* [#10105](https://github.com/redisearch/redisearch/pull/10105) FT.SEARCH and FT.AGGREGATE return Unknown argument errors on 2.8 shards during rolling upgrades when a newer coordinator injects internal query arguments. (MOD-16047) \n* [#9579](https://github.com/redisearch/redisearch/pull/9579) Server becomes unresponsive when FT.CREATE is called with an extremely large number of fields. (MOD-6411) \n* [#9825](https://github.com/redisearch/redisearch/pull/9825) FT.SYNUPDATE leaves the synonym map partially mutated when an update batch exceeds synonym or group-ID limits. (MOD-15402) \n* [#9534](https://github.com/redisearch/redisearch/pull/9534) FT.INFO num_records is skewed for indexes with vector fields because vector entries are never decremented on deletion. (MOD-15487)" }, { "id": "v2636-may-2026", "title": "v2.6.36 (May 2026)", "role": "content", "text": "This is a maintenance release for Redis Search 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\nBug fixes:\n\n- [#Q8950](https://github.com/redisearch/redisearch/pull/8950) `FT.CURSOR` enters an infinite loop when the invoking user's ACL restricts access to the cursor's index, causing CPU exhaustion and shard restarts. (MOD-14479)\n- [#Q8468](https://github.com/redisearch/redisearch/pull/8468) Garbage collector mishandles out-of-memory conditions on replicas, causing instability under memory pressure during indexing. (MOD-14066)\n- [#Q6792](https://github.com/redisearch/redisearch/pull/6792) Counter result-processor cleanup uses an incorrect type cast, risking memory corruption when freeing query pipeline resources. (MOD-11216)\n- [#Q6998](https://github.com/redisearch/redisearch/pull/6998) `FT.INFO` is fanned out to replicas in clustered deployments, producing inconsistent responses and unnecessary inter-node traffic.\n- [#Q6871](https://github.com/redisearch/redisearch/pull/6871) Coupled grow/shrink logic in Flat and HNSW vector index containers causes excess memory churn for large or fluctuating vector datasets. (MOD-10559)" }, { "id": "v2632-august-2025", "title": "v2.6.32 (August 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\nBug fixes:\n- [#6405](https://github.com/redisearch/redisearch/pull/6405) Validate compatibility against RedisJSON version upon open key (MOD-10298).\n- [#6419](https://github.com/redisearch/redisearch/pull/6419) Errors when loading schema from RDB get incorrectly cleared (MOD-10307).\n- [#6604](https://github.com/redisearch/redisearch/pull/6604) `FLUSHDB` while active queries are still running could lead to a crash due to premature release of the CURSOR (MOD-10681).\n\nImprovements:\n- [#6467](https://github.com/redisearch/redisearch/pull/6467) Handle excessive error logs when handling JSON.DEL errors (MOD-10266).\n- [#6659](https://github.com/redisearch/redisearch/pull/6659) Time measurement on `FT.PROFILE` using thread-independent clock mechanism (MOD-10622)." }, { "id": "v2631-june-2025", "title": "v2.6.31 (June 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH`: There is a critical bug that may affect a subset of users. Upgrade!\n\nBug fixes:\n- [#6349](https://github.com/redisearch/redisearch/pull/6349) Search on terms larger than 128 characters could lead to missing matches (MOD-6786).\n- [#6305](https://github.com/redisearch/redisearch/pull/6305) While iterating over a large index, frequent document updates could hit the `TIMEOUT`, causing a crash (MOD-9856).\n- [#6191](https://github.com/redisearch/redisearch/pull/6191) Reindexing from an RDB with multiple vector indexes could lead to a crash because of the cluster health check - NodeWD (MOD-9220)." }, { "id": "v2630-may-2025", "title": "v2.6.30 (May 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\nBug fixes:\n- [#6032](https://github.com/redisearch/redisearch/pull/6032) `FT.CURSOR...DEL` while another thread is reading it could lead to a crash (MOD-9408,MOD-9432,MOD-9433,MOD-9434,MOD-9435)\n- [#5965](https://github.com/redisearch/redisearch/pull/5965) Indexing documents using `TEXT` without the text in the documents leads to an `inf` or `nan` score (MOD-9423)\n- [#6058](https://github.com/redisearch/redisearch/pull/6058) Avoid lazy expiration in background indexing for Active-Active setup, preventing keys from expiring incorrectly (MOD-9486)\n- [#5962](https://github.com/redisearch/redisearch/pull/5962) Empty results with RESP3 due to the `TIMEOUT` even if setting to deliver partial results using the `ON_TIMEOUT` policy (MOD-8482)\n- [#5962](https://github.com/redisearch/redisearch/pull/5962) Cursor with RESP3 on `FT.AGGREGATE` is never depleted, blocking queries if the cursor limit is achieved (MOD-8515)\n- [#5962](https://github.com/redisearch/redisearch/pull/5962) Using `FT.CURSOR READ` on queries that timed out led to fewer results than expected (MOD-8606)\n\nImprovements:\n- [#6009](https://github.com/redisearch/redisearch/pull/6009) Parser for intersections on parentheses and sub-queries order won't affect full-text scores (MOD-9278)\n- [#5962](https://github.com/redisearch/redisearch/pull/5962) Fixed a coordinator race condition preventing the premature release and avoiding errors and inconsistencies during query executions (MOD-8794)" }, { "id": "v2629-april-2025", "title": "v2.6.29 (April 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `LOW` No need to upgrade unless there are new features you want to use.\n\nImprovements:\n- [#5940](https://github.com/redisearch/redisearch/pull/5940) Improved performance (reduced CPU time) of collecting vector index statistics (MOD-9354)\n- [#5816](https://github.com/redisearch/redisearch/pull/5816) Improved accuracy of index memory reporting by correcting a bug that caused negative memory counts (MOD-5904)" }, { "id": "v2628-march-2025", "title": "v2.6.28 (March 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `LOW` No need to upgrade unless there are new features you want to use.\n\nBug fixes:\n- [#5712](https://github.com/redisearch/redisearch/pull/5712) Weights in the query string are ignored if using `SCORER BM25` (MOD-7896)" }, { "id": "v2627-february-2025", "title": "v2.6.27 (February 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `LOW` No need to upgrade unless there are new features you want to use.\n\nBug fixes:\n\n- [#5648](https://github.com/redisearch/redisearch/pull/5648) `FT.SEARCH` using Cyrillic characters and wildcards delivering no results (MOD-7944)" }, { "id": "v2626-february-2025", "title": "v2.6.26 (February 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\nBug fixes:\n- [#5606](https://github.com/redisearch/redisearch/pull/5606) Changes on the memory block reading logic could cause crash on `FT.SEARCH` with error \"_Redis 7.4.2 crashed by signal: 11, si_code: 128_\"\n\nKnown limitations:\n- Only the first 128 characters of string fields are normalized to lowercase during ingestion (for example, on `HSET`).\n Example:\n\n [code example]\n\n The `FT.SEARCH` command will return no documents." }, { "id": "v2625-january-2025", "title": "v2.6.25 (January 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\nBug fixes:\n- [#5484](https://github.com/redisearch/redisearch/pull/5484) NOSTEM option does not work on query, just tokenising (MOD-7634)\n- [#5543](https://github.com/redisearch/redisearch/pull/5543) Querying for the latest document added to the index may result in a crash if the last block is not read (MOD-8561)." }, { "id": "v2624-january-2025", "title": "v2.6.24 (January 2025)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `SECURITY`: There are security fixes in the release.\n\n- **Security and privacy:**\n - [#5458](https://github.com/redisearch/redisearch/pull/5458) (CVE-2024-51737) Query: potential out-of-bounds write (MOD-8486)\n\n- Bug fixes:\n - [#5302](https://github.com/redisearch/redisearch/pull/5302) Prefix/Infix/Suffix queries longer than 1024 chars could cause a crash (MOD-7882)\n - [#5281](https://github.com/redisearch/redisearch/pull/5281) `FT.CURSOR READ` retrieving deleted `TAG` fields cause a crash (MOD-8011)\n - [#5168](https://github.com/redisearch/redisearch/pull/5168) Cursors from queries that timed out weren't depleted causing exhaustion of number of cursors available(MOD-8009)" }, { "id": "v2621-august-2024", "title": "v2.6.21 (August 2024)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\n- Bug fixes:\n - [#4944](https://github.com/redisearch/redisearch/pull/4944) Adjusting the module configuration to avoid routing overload on the first shard in a clustered database (MOD-7505)\n - [#4897](https://github.com/redisearch/redisearch/pull/4897) - `FT.AGGREGATE` with `VERBATIM` option is not handled by the shards in cluster mode (MOD-7463)\n - [#4918](https://github.com/redisearch/redisearch/pull/4918) - Union query, similar to `\"is|the\"`, starting with 2 [stopwords](https://redis.io/docs/latest/develop/ai/search-and-query/advanced-concepts/stopwords/) could cause a crash (MOD-7495)\n - [#4919](https://github.com/redisearch/redisearch/pull/4919) - Counting twice the field statistics at the `#search` section of an `INFO` response (MOD-7339)\n - [#4923](https://github.com/redisearch/redisearch/pull/4923) - Loop when using the wildcard `w'term'` and prefix/infix/suffix pattern `'ter*'`, causing the shard to restart (MOD-7453)\n - [#4954](https://github.com/redisearch/redisearch/pull/4954) `FT.PROFILE` on `AGGREGATE` numeric queries could cause a crash due to reusing the internal `CURSOR` in a large range of numeric values (MOD-7454)" }, { "id": "v2620-july-2024", "title": "v2.6.20 (July 2024)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `MODERATE`: Program an upgrade of the server, but it's not urgent.\n\n- Bug fixes:\n - [#4755](https://github.com/RediSearch/RediSearch/pull/4755) - Correct return the maximum value for negative values when using `MAX` reducer (MOD-7252)\n - [#4733](https://github.com/RediSearch/RediSearch/pull/4733) - Separators ignored when escaping backslash `\\` after the escaped character such as in `hello\\\\,world` ignoring `,` (MOD-7240)\n - [#4717](https://github.com/RediSearch/RediSearch/pull/4717) - Sorting by multiple fields as in `SORTBY 2 @field1 @field2` was ignoring the subsequent field (MOD-7206)\n\n- Improvements:\n - [#4793](https://github.com/RediSearch/RediSearch/pull/4793) - Add character validations to simple string replies and escape it when required(MOD-7258)\n - [#4769](https://github.com/RediSearch/RediSearch/pull/4769) - Indicate which value is missing on the error message at the aggregation pipeline (MOD-7201)\n - [#4746](https://github.com/RediSearch/RediSearch/pull/4746) - `GROUPBY` recursion cleanup (MOD-7245)" }, { "id": "v2619-june-2024", "title": "v2.6.19 (June 2024)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `MODERATE`: Program an upgrade of the server, but it's not urgent.\n\n- Bug fixes:\n\n - [#4616](https://github.com/RediSearch/RediSearch/pull/4616) Shards become unresponsive when using FT.AGGREGATE with APPLY 'split(...)' (MOD-6759)\n - [#4557](https://github.com/RediSearch/RediSearch/pull/4557) FT.EXPLAIN returns additional } when querying using wildcards (MOD-6768)\n - [#4647](https://github.com/RediSearch/RediSearch/pull/4647) FT.DROPINDEX with DD flag deleted keys in one AA cluster but not the others (MOD-1855)\n\n- Improvements:\n\n - [#4599](https://github.com/RediSearch/RediSearch/pull/4599) Report additional memory consumed by the TAG and TEXT tries (MOD-5902)\n - [#4688](https://github.com/RediSearch/RediSearch/pull/4688) Add missing FT.INFO fields when used within a cluster (MOD-6920)" }, { "id": "v2618-april-2024", "title": "v2.6.18 (April 2024)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `LOW`: No need to upgrade unless there are new features you want to use.\n\nDetails:\n\n- Bug fixes:\n\n - [#4557](https://github.com/RediSearch/RediSearch/pull/4557) Additional \"`}`\" on wildcards replies for `FT.EXPLAIN` (MOD-6768)" }, { "id": "v2617-april-2024", "title": "v2.6.17 (April 2024)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH`: There is a critical bug that may affect a subset of users. Upgrade!\n\nDetails:\n\n- Bug fixes:\n\n - [#4524](https://github.com/RediSearch/RediSearch/pull/4524) `FT.CURSOR READ` in a numeric query causing a crash (MOD-6597)\n - [#4543](https://github.com/RediSearch/RediSearch/pull/4543) `FT.SEARCH` accessing an inexistent memory address causes a crash if using deprecated `FT.ADD` command (MOD-6599)\n - [#4535](https://github.com/RediSearch/RediSearch/pull/4535) `FT.PROFILE` with incorrect arguments could cause a crash on cluster setup (MOD-6791)\n - [#4540](https://github.com/RediSearch/RediSearch/pull/4540) Unfree memory from an existing RDB while re-indexing loading a new RDB could cause a crash (MOD-6831, 6810)\n - [#4485](https://github.com/RediSearch/RediSearch/pull/4485) Some parameter settings using just prefixes instead of full values were working (MOD-6709)\n\n- Improvements:\n\n - [#4502](https://github.com/RediSearch/RediSearch/pull/4502) Handle error properly when trying to execute Search commands on cluster setup as part of `MULTI ... EXEC` or LUA script (MOD-6541)" }, { "id": "v2616-march-2024", "title": "v2.6.16 (March 2024)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `MODERATE` : Program an upgrade of the server, but it's not urgent.\n\nDetails:\n\n- Bug fixes:\n\n - [#4477](https://github.com/RediSearch/RediSearch/pull/4477) Split `INFIX` and `SUFFIX` report on `FT.EXPLAIN` and `FT.EXPLAINCLI` (MOD-6186)\n - [#4468](https://github.com/RediSearch/RediSearch/pull/4468) Memory leak upon suffix query for a `TAG` indexed with `WITHSUFFIXTRIE` (MOD-6644)\n - [#4407](https://github.com/RediSearch/RediSearch/pull/4407) Clustered `FT.SEARCH` hangs forever without replying when an invalid topology is found (MOD-6557)\n - [#4359](https://github.com/RediSearch/RediSearch/pull/4359) Searching for a synonym will iterate in the same group multiple times, causing a performance hit (MOD-6490)\n - [#4310](https://github.com/RediSearch/RediSearch/pull/4310) Memory tracking on cluster setups causing high memory usage and potentially Out-of-Memory (MOD-6123, MOD-5639)" }, { "id": "v2615-december-2023", "title": "v2.6.15 (December 2023)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH` : There is a critical bug that may affect a subset of users. Upgrade!\n\nDetails:\n\n- Bug fixes:\n\n - [#4244](https://github.com/RediSearch/RediSearch/pull/4244), [#4255](https://github.com/RediSearch/RediSearch/pull/4255) Profiling `FT.AGGREGATE` using the `WITHCURSOR` flag causes a crash due to timeout (MOD-5512)\n - [#4238](https://github.com/RediSearch/RediSearch/pull/4238) Memory excessively growing on databases caused by unbalanced nodes on inverted index trie (MOD-5880, MOD-5952, MOD-6003) \n - [#3995](https://github.com/RediSearch/RediSearch/pull/3995) `FT.CURSOR READ` with geo queries causing a crash when data is updated between the cursor reads (MOD-5646) \n - [#4155](https://github.com/RediSearch/RediSearch/pull/4155) `FT.SEARCH` not responding when using TLS encryption on Amazon Linux 2 (MOD-6012)\n\n- Improvements:\n\n - [#4176](https://github.com/RediSearch/RediSearch/pull/4176) Initialization of the maximum numeric value range leading to a better balance of the index leaf splitting (MOD-6232) \n - [#4123](https://github.com/RediSearch/RediSearch/pull/4123) Possibly problematic index name alias check-in command multiplexing (MOD-5945)\n - [#4195](https://github.com/RediSearch/RediSearch/pull/4195) Query optimization when predicate contains multiple `INTERSECTION` (AND) of `UNION` (OR) (MOD-5910)" }, { "id": "v2614-november-2023", "title": "v2.6.14 (November 2023)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `SECURITY`: There are security fixes in the release.\n\nDetails:\n\n- Bug fixes:\n\n - [#3783](https://github.com/RediSearch/RediSearch/pull/3783) Broken lower and upper `APPLY` functions in `FT.AGGREGATE` on `DIALECT 3` (MOD-5041)\n - [#3823](https://github.com/RediSearch/RediSearch/pull/3823) `APPLY` or `FILTER` expression causing a leak (MOD-5751)\n - [#3899](https://github.com/RediSearch/RediSearch/pull/3899) Connection using TLS fail on Redis (MOD-5768)\n - [#3910](https://github.com/RediSearch/RediSearch/pull/3910) Heavy document updates causing memory growth if memory blocks weren't properly released (MOD-5181)(MOD-5757)\n - [#3928](https://github.com/RediSearch/RediSearch/pull/3928) Queries with `WITHCURSOR` making memory growth since `CURSOR` wasn't invalidated in the shards (MOD-5580)\n - [#3946](https://github.com/RediSearch/RediSearch/pull/3946) Vector range query could cause Out-of-Memory due to memory corruption (MOD-5791)\n - [#3972](https://github.com/RediSearch/RediSearch/pull/3972) Adding new nodes to OSS cluster can cause a crash (MOD-5778)\n - [#3957](https://github.com/RediSearch/RediSearch/pull/3957) After cleaning the index, the GC could corrupt unique values (MOD-5815)\n - [#4002](https://github.com/RediSearch/RediSearch/pull/4002) Setting a low `MAXIDLE` parameter value in `FT.AGGREGATE` causes a crash (MOD-5608)\n\n- Security and privacy:\n\n - [#3844](https://github.com/RediSearch/RediSearch/pull/3844) Limits maximum phonetic length to avoid vulnerability (MOD 5767)" }, { "id": "v2612-july-2023", "title": "v2.6.12 (July 2023)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH`: There is a critical bug that may affect a subset of users. Upgrade!\n\nDetails:\n\n- Bug fixes:\n\n - [#3557](https://github.com/RediSearch/RediSearch/pull/3557) `TIMEOUT` configuration on `FT.AGGREGATE` query being ignored (MOD-5208)\n - [#3552](https://github.com/RediSearch/RediSearch/pull/3552) `FT.CURSOR READ` on `JSON` numeric queries not returning results (MOD-4830)\n - [#3606](https://github.com/RediSearch/RediSearch/pull/3606) Update numeric inverted index `numEntries` avoiding excessive memory consumption (MOD-5181)\n - [#3597](https://github.com/RediSearch/RediSearch/pull/3597) Duplicating alias as output name on `FT.AGGREGATE` reducer (`REDUCE` argument) doesn't return results (MOD-5268)\n - [#3654](https://github.com/RediSearch/RediSearch/pull/3654) Added check for `@` prefix on `GROUPBY` fields returning an error instead of wrong results\n\n- Improvements:\n\n - [#3628](https://github.com/RediSearch/RediSearch/pull/3628) Background indexing scanning performance (MOD-5259)\n - [#3259](https://github.com/RediSearch/RediSearch/pull/3259) Allow alias name beginning with `as`\n - [#3641](https://github.com/RediSearch/RediSearch/pull/3641) Indexing sanitizing trigger in heavy data updates scenario" }, { "id": "v269-april-2023", "title": "v2.6.9 (April 2023)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH`: There is a critical bug that may affect a subset of users. Upgrade!\n\nDetails:\n\n- Bug fixes:\n\n - [#3468](https://github.com/RediSearch/RediSearch/pull/3468) KNN searching for 0 vectors with a filter resulted in crash (MOD-5006)\n - [#3499](https://github.com/RediSearch/RediSearch/pull/3499) `MAXSEARCHRESULTS` set to `0` causing `FT.SEARCH` to crash (MOD-5062)\n - [#3494](https://github.com/RediSearch/RediSearch/pull/3494) Removing `MAXSEARCHRESULTS` limit causes crash on `FT.AGGREGATE` (MOD-4974)\n - [#3504](https://github.com/RediSearch/RediSearch/pull/3504) Uninitialised vector similarity query parameter bug (MOD-5063)\n\n- Improvements:\n\n - [#3430](https://github.com/RediSearch/RediSearch/pull/3430) Improve min-max heap structure for better readability and performance\n - [#3450](https://github.com/RediSearch/RediSearch/pull/3450) Display `NOHL` option in `FT.INFO` command\n - [#3534](https://github.com/RediSearch/RediSearch/pull/3534) Vector Similarity 0.6.1: Improve multi-value index deletion logic ([#346](https://github.com/RedisAI/VectorSimilarity/pull/346))" }, { "id": "v266-march-2023", "title": "v2.6.6 (March 2023)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `MODERATE`: Program an upgrade of the server, but it's not urgent.\n\nDetails:\n\n- Bug fixes:\n\n - [#3403](https://github.com/RediSearch/RediSearch/pull/3403) Fix suffix and prefix matching when using `CASESENSITIVE` flag (MOD-4872)\n\n- Improvements:\n\n - [#3397](https://github.com/RediSearch/RediSearch/pull/3397) Improve the Vecsim initial capacity default value" }, { "id": "v265-february-2023", "title": "v2.6.5 (February 2023)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `HIGH`: There is a critical bug that may affect a subset of users. Upgrade!\n\nDetails:\n\n- Bug fixes:\n\n - [#3354](https://github.com/RediSearch/RediSearch/pull/3354) Library update preventing a crash during cluster failover (MOD-4560)\n - [#3357](https://github.com/RediSearch/RediSearch/pull/3357) Handling division by zero in expressions preventing nodes to restart (MOD-4296)\n - [#3332](https://github.com/RediSearch/RediSearch/pull/3332) Fix wildcards `*` queries on `DIALECT 2` and `DIALECT 3`\n\n- Improvements:\n \n - [#3361](https://github.com/RediSearch/RediSearch/pull/3361) Enable the use of IPv6 for all cluster and module communication" }, { "id": "v264-december-2022", "title": "v2.6.4 (December 2022)", "role": "content", "text": "This is a maintenance release for RediSearch 2.6.\n\nUpdate urgency: `MODERATE`: Program an upgrade of the server, but it's not urgent.\n\nDetails:\n\n- Bug fixes:\n\n - [#3289](https://github.com/RediSearch/RediSearch/pull/3289) Potential crash when querying multiple fields (MOD-4639)\n - [#3279](https://github.com/RediSearch/RediSearch/pull/3279) Potential crash when querying using wildcard `*` on TAG field (MOD-4653)\n\n- Improvements:\n \n - [#3256](https://github.com/RediSearch/RediSearch/pull/3256) Support IPv6 on cluster set command\n - [#3194](https://github.com/RediSearch/RediSearch/pull/3194) Add the query dialects that are in use to `FT.INFO` and `INFO MODULE` commands (MOD-4232)\n - [#3258](https://github.com/RediSearch/RediSearch/pull/3258) Add the module version and Redis version to `INFO MODULE`" }, { "id": "v26-ga-v263-november-2022", "title": "v2.6 GA (v2.6.3) (November 2022)", "role": "content", "text": "This is the General Availability release of RediSearch 2.6." }, { "id": "highlights", "title": "Highlights", "role": "content", "text": "This new major version introduces the ability to search using **wildcard queries** for TEXT and TAG fields. This enables the frequently requested feature **suffix search** (`*vatore` and `ant?rez` are now supported).\nIn addition, the 2.6 release is all about **multi-value indexing and querying of attributes** for any attribute type ( [Text](https://redis.io/docs/latest/develop/ai/search-and-query/indexing#index-json-arrays-as-text), [Tag](https://redis.io/docs/latest/develop/ai/search-and-query/indexing#index-json-arrays-as-tag), [Numeric](https://redis.io/docs/latest/develop/ai/search-and-query/indexing#index-json-arrays-as-numeric), [Geo](https://redis.io/docs/latest/develop/ai/search-and-query/indexing#index-json-arrays-as-geo) and [Vector](https://redis.io/docs/latest/develop/ai/search-and-query/indexing#index-json-arrays-as-vector)) defined by a [JSONPath](https://redis.io/docs/latest/develop/data-types/json/path) leading to an array or to multiple scalar values.\nLastly, this version adds support for indexing double-precision floating-point vectors and range queries from a given vector." }, { "id": "whats-new-in-26", "title": "What's new in 2.6", "role": "content", "text": "" }, { "id": "details", "title": "Details", "role": "content", "text": "- Improvements:\n\n - [#2886](https://github.com/RediSearch/RediSearch/pull/2886) Support for [wildcard queries](https://redis.io/docs/latest/develop/ai/search-and-query/advanced-concepts/query_syntax#wildcard-matching) for TEXT and TAG fields, where\n - `?` matches any single character\n - `*` matches zero or more characters\n - use `'` and `\\` for escaping, other special characters are ignored\n - [#2932](https://github.com/RediSearch/RediSearch/pull/2932) Optimized wildcard query support (i.e., suffix trie)\n - Multi-value indexing and querying\n - [#2819](https://github.com/RediSearch/RediSearch/pull/2819), [#2947](https://github.com/RediSearch/RediSearch/pull/2947) Multi-value text search - perform full-text search on an [array of strings or on a JSONPath](https://redis.io/docs/latest/develop/ai/search-and-query/indexing#index-json-arrays-as-tag) leading to multiple strings\n - [#3131](https://github.com/RediSearch/RediSearch/pull/3131) Geo [#3118](https://github.com/RediSearch/RediSearch/pull/3118) Vector [#2985](https://github.com/RediSearch/RediSearch/pull/2985) Numeric [#3180](https://github.com/RediSearch/RediSearch/pull/3180) Tag\n - [#3060](https://github.com/RediSearch/RediSearch/pull/3060) Return JSON rather than scalars from multi-value attributes. This is enabled via Dialect 3 in order not to break existing applications.\n - Support indexing and querying of multi-value JSONPath attributes and/or arrays (requires JSON >2.4.1)\n - [#3182](https://github.com/RediSearch/RediSearch/pull/3182) Support for `SORTABLE` fields on JSON in an implicit un-normalized form (UNF)\n - [#3156](https://github.com/RediSearch/RediSearch/pull/3156) Vector similarity 0.5.1:\n - Better space optimization selection ([#175](https://github.com/RedisAI/VectorSimilarity/pull/175)) \n - Aligning index capacity with block size ([#177](https://github.com/RedisAI/VectorSimilarity/pull/177)) \n - [#3129](https://github.com/RediSearch/RediSearch/pull/3129) Support FLOAT64 as vector data type \n - [#3176](https://github.com/RediSearch/RediSearch/pull/3176) Range query support\n - [#3105](https://github.com/RediSearch/RediSearch/pull/3105) Support query attributes for vector queries\n\n- Bugs (since 2.6-RC1 / v2.6.1):\n\n - [#3197](https://github.com/RediSearch/RediSearch/pull/3197) Failure to create temporary indices\n - [#3098](https://github.com/RediSearch/RediSearch/pull/3098) Wrong return value in Geo query\n - [#3230](https://github.com/RediSearch/RediSearch/issues/3230) Use the correct total number of matching records\n\n> [!NOTE]\n> With this release, we stop supporting direct upgrades from RediSearch v1.4 and v1.6 that are End-of-Life. Such RDB files can still be upgraded to RediSearch 2.0 first.\n\n> [!NOTE]\n> If indexing and querying RedisJSON data structures, this version is best combined with RedisJSON 2.4 GA (v2.4.1 onwards)." } ], "examples": [ { "id": "v2626-february-2025-ex0", "language": "plaintext", "code": "HSET doc __score 1.0 name \"idx1S...S\" mynum 1 # Assume \"S...S\" is a string of 252 capital S's\n FT.CREATE \"idx\" SCHEMA \"name\" \"TEXT\" \"mynum\" \"NUMERIC\"\n FT.SEARCH \"idx\" \"@name:idx1S...S\" # Assume \"S...S\" is a string of 252 capital S's", "section_id": "v2626-february-2025" } ] }