What's Changed

• Added support for index pushdown for sort queries filtering on _id field.
• Validate $jsonSchema syntax during rule creation or modification(schema validation) [Preview]. Requires documentdb.enableSchemaValidation GUC to be on.
• Support schema enforcement with CSFLE integration [Preview]. Requires the GUC documentdb.enableSchemaEnforcementForCSFLE (enabled by default) and documentdb.enableSchemaValidation to be set to on.
• TTL index uses ordered index scan via index hints. TTL index performs batch deletions continuously for up to 60 seconds, instead of once every 60 seconds.

Full Changelog: v0.106-0...v0.107-0

What's Changed

This release focuses on expanding indexing capabilities, enhancing role visibility, and improving query planning performance, while also delivering important bug fixes for stability and compatibility.

• Introduced support for index hints, giving developers the option to explicitly specify an index during query execution
• Enabled index-only scans on ordered indexes for count queries where filters can be pushed down to the index and no projections are applied.
• Enabled parallel index builds to improve indexing speed and efficiency.
• Add support to prefer new ordered index over existing non-ordered indexes.
• Improvised the point lookup queries with better cost estimates to ensure that we use primary key index.
• Enhanced usersInfo and connectionStatus commands to return all supported roles (ReadWrite, UserAdmin, Root) and privilege sets (dbAdmin, userAdmin, clusterMonitor, clusterManager, hostManager).
• Improved query planning performance by introducing a custom execution plan for insert operations.

Bug fixes

• Fix bug in $lastN and $bottomN
• Fix Response type of update to add compatibility to C++ drivers requiring response in 32 bit .
• Enhanced error messages for output stages $merge and $out

Full Changelog: v0.105-0...v0.106-0

What's Changed

1. Support $bucketAuto aggregation stage, with multiple granularity types.
2. Support connectionStatus command.

Full Changelog: v0.104-0...v0.105-0

This release adds collation support for key operators, enables background index builds, improves user CRUD defaults, and expands feature/config coverage.

1. Add string case support for $toDate operator.
2. Support sort with collation in runtime.
3. Support collation with $indexOfArray aggregation operator.
4. Support collation with array and object comparisons.
5. Default support enabled for Read & Read\Write privileges with secondary users.
6. Enable $let support for delete queries (requires EnableVariablesSupportForWriteCommands to be enabled).
7. Support compact command (requires documentdb.enablecompact GUC).
8. Enable role privileges for usersInfo command.

Performance
Enable rum_enable_index_scan as default.

Tooling / Infra
Add public documentdb-local Docker image with gateway to GHCR.

Full Changelog: v0.103-0...v0.104-0

This release adds collation support across aggregation and find on sharded collections, enhances vector indexing with 4,000 dimension support, introduces binData conversions, ARM64 Docker builds (preview), documentdb gateway, and fixes database listing for >2GB.

1. Support collation with aggregation and find on sharded collections [Feature]
2. Support $convert on binData to binData, string to binData and binData to string (except with format: auto) [Feature]
3. Support half-precision vector indexing, vectors can have up to 4,000 dimensions [Feature]
4. Support collation with $documents and $replaceWith stage of the aggregation pipeline [Feature]
5. Push pg_documentdb_gw for documentdb connections [Feature]
6. Support ARM64 architecture when building docker container [Preview]
7. Fix list_databases for databases with size > 2 GB [Bugfix] (#119)

Full Changelog: v0.102-0...v0.103-0

New Release with improvements to vector queries, adding support for collation and several commands and more...

1. Support index pushdown for vector search queries [Bugfix]
2. Support exact search for vector search queries [Feature]
3. Inline $match with let in $lookup pipelines as JOIN Filter [Perf]
4. Support TTL indexes [Bugfix] (#34)
5. Support joining between postgres and documentdb tables [Feature] (#61)
6. Support current_op command [Feature] (#59)
7. Support for list_databases command [Feature] (#45)
8. Disable analyze statistics for unique index uuid columns which improves resource usage [Perf]
9. Support collation with $expr, $in, $cmp, $eq, $ne, $lt, $lte, $gt, $gte comparison operators (Opt-in) [Feature]
10. Support collation in find, aggregation $project, $redact, $set, $addFields, $replaceRoot stages (Opt-in) [Feature]
11. Support collation with $setEquals, $setUnion, $setIntersection, $setDifference, $setIsSubet in the aggregation pipeline (Opt-in) [Feature]
12. Support unique index truncation by default with new operator class [Feature]
13. Top level aggregate command let variables support for $geoNear stage [Feature]
14. Enable Backend Command support for Statement Timeout [Feature]
15. Support type aggregation operator $toUUID. [Feature]
16. Support Partial filter pushdown for $in predicates [Perf]
17. Support the $dateFromString operator with full functionality [Feature]
18. Support extended syntax for $getField aggregation operator. Now the value of 'field' could be an expression that resolves to a string. [Feature]

Full change log: v0.101-0...v0.102-0

See what's new in DocumentDB:

1. Push $graphlookup recursive CTE JOIN filters to index [Perf]
2. Build pg_documentdb for PostgreSQL 17 [Infra] (#13)
3. Enable support of currentOp aggregation stage, along with collstats, dbstats, and indexStats [Commands] (#52)
4. Allow inlining $unwind with $lookup with preserveNullAndEmptyArrays [Perf]
5. Skip loading documents if group expression is constant [Perf]
6. Fix Merge stage not outputing to target collection [Bugfix] (#20)

Full change log: v0.100-0...v0.101-0

Initial version of DocumentDB includes support for

1. CRUD, Query and aggregation operations on BSON datatype.
2. Indexing the BSON field.
3. Compatible with pg_vector to support vector queries.
4. Compatible with postgis for geospatial queries.
5. Compatible with RUM indexes for Full text searches.