ScyllaDB Docs ScyllaDB Manual CQL Reference Global Secondary Indexes

Caution

You're viewing documentation for a previous version of ScyllaDB Manual. Switch to the latest stable version.

Global Secondary Indexes¶

CQL supports creating secondary indexes on tables, allowing queries on the table to use those indexes. A secondary index is identified by a name defined by:

index_name: re('[a-zA-Z_0-9]+')

CREATE INDEX¶

Creating a secondary index on a table uses the CREATE INDEX statement:

create_index_statement: CREATE [ CUSTOM ] INDEX [ IF NOT EXISTS ] [ `index_name` ]
                      :     ON `table_name` '(' `index_identifier` ')'
                      :     [ USING `string` [ WITH OPTIONS = `map_literal` ] ]
index_identifier: `column_name`
                :| ( FULL ) '(' `column_name` ')'

For instance:

CREATE INDEX userIndex ON NerdMovies (user);
CREATE INDEX ON Mutants (abilityId);

The CREATE INDEX statement is used to create a new (automatic) secondary index for a given (existing) column in a given table. A name for the index itself can be specified before the ON keyword, if desired. If data already exists for the column, it will be indexed asynchronously. After the index is created, new data for the column is indexed automatically at insertion time.

Local Secondary Index¶

Local Secondary Indexes is an enhancement of Global Secondary Indexes, which allows ScyllaDB to optimize the use case in which the partition key of the base table is also the partition key of the index. Local Secondary Index syntax is the same as above, with extra parentheses on the partition key.

index_identifier: `column_name`
                :| ( PK ) | KEYS | VALUES | FULL ) '(' `column_name` ')'

Example:

CREATE TABLE menus (location text, name text, price float, dish_type text, PRIMARY KEY(location, name));
CREATE INDEX ON menus((location),dish_type);

Vector Index Experimental ScyllaDB Cloud¶

Note

Vector indexes are supported in ScyllaDB Cloud only in the clusters that have the vector search feature enabled. Moreover, vector indexes are an experimental feature that:

is not suitable for production use,

does not guarantee backward compatibility between ScyllaDB versions,

does not support all the features of ScyllaDB (e.g., tracing, filtering, TTL).

ScyllaDB supports creating vector indexes on tables, allowing queries on the table to use those indexes for efficient similarity search on vector data.

The vector index is the only custom type index supported in ScyllaDB. It is created using the CUSTOM keyword and specifying the index type as vector_index. Example:

CREATE CUSTOM INDEX vectorIndex ON ImageEmbeddings (embedding)
USING 'vector_index'
WITH OPTIONS = {'similarity_function': 'COSINE', 'maximum_node_connections': '16'};

The following options are supported for vector indexes. All of them are optional.

Option Name	Description	Default Value
`similarity_function`	The similarity function to use for vector comparisons. Supported values are: `COSINE`, `EUCLIDEAN`, and `DOT_PRODUCT`. `DOT_PRODUCT` requires vectors to be normalized, meaning each vector should have unit length (L2 norm = 1). For more information, see Vector normalization.	`COSINE`
`maximum_node_connections`	The maximum number of connections per node in the HNSW graph. In other HNSW implementations it is often denoted as `m`. Higher values lead to better recall (i.e., more relevant results are found) but increase memory usage and index size. Supported values are integers between 1 and 512.	`16`
`construction_beam_width`	The beam width to use during index construction. In other HNSW implementations it is often denoted as `efConstruction`. Higher values lead to better recall (i.e., more relevant results are found) but increase index creation time and memory usage. Supported values are integers between 1 and 4096.	`128`
`search_beam_width`	The beam width to use during index search. In other HNSW implementations it is often denoted as `efSearch`. Higher values lead to better recall (i.e., more relevant results are found) but increase query latency. Supported values are integers between 1 and 4096.	`64`

DROP INDEX¶

Dropping a secondary index uses the DROP INDEX statement:

drop_index_statement: DROP INDEX [ IF EXISTS ] `index_name`

The DROP INDEX statement is used to drop an existing secondary index. The argument of the statement is the index name, which may optionally specify the keyspace of the index.

Additional Information¶

The following courses are available from ScyllaDB University:

Apache®, Apache Cassandra®, Cassandra®, the Apache feather logo and the Apache Cassandra® Eye logo are either registered trademarks or trademarks of the Apache Software Foundation in the United States and/or other countries. No endorsement by The Apache Software Foundation is implied by the use of these marks.

Was this page helpful?