New in 3.2
|
Couchbase Lite 3.0 introduces some breaking changes. If you’re upgrading from 2.x, refer to the appropriate upgrade page — see: Upgrading Users should be able to upgrade to 3.1.x from 3.0.x without manual intervention. |
Release 3.2.1 (November 2024)
New Features
Array UNNEST and the Array Index
You can use UNNEST in queries to unpack arrays within a document into individual rows. This capability makes it possible to join them with their parent object in the query.
You can use UNNEST within the FROM clause. You can chain UNNEST to perform multi-level UNNEST.
You can also use a new type of index, the Array Index, to allow querying with UNNEST more efficiently.
For more information about Array UNNEST, see:
For more information about Array indexes, see:
Release 3.2.0 (August 2024)
| Databases upgraded from 3.1.x to 3.2.x cannot be downgraded. |
New Features
Vector Search
|
Vector Search is available only for 64-bit architectures and Intel processors that support the Advanced Vector Extensions 2 (AVX2) instruction set. To verify whether your device supports the AVX2 instructions set, follow these instructions. |
Vector Search is now available on Couchbase Lite for all platforms. Vector Search is a sophisticated data retrieval technique that focuses on matching the contextual meanings of search queries and data entries, rather than simple text matching. Vectors are represented by arrays of numbers known as embeddings, which are generated by Large Language Models (LLMs) to represent objects such as text, images, and audio. You can use Vector Search to efficiently find similar items or content based on the similarity of their vector representations. This is useful for reducing the cost per query, performing semantic or similarity search, providing recommendations among others.
Read more at: