Nóta
Teastaíonn údarú chun rochtain a fháil ar an leathanach seo. Is féidir leat triail a bhaint as shíniú isteach nó eolairí a athrú.
Teastaíonn údarú chun rochtain a fháil ar an leathanach seo. Is féidir leat triail a bhaint as eolairí a athrú.
Note
Databricks CLI use is subject to the Databricks License and Databricks Privacy Notice, including any Usage Data provisions.
The vector-search-indexes command group within the Databricks CLI contains commands to manage vector search indexes. A vector search index is an efficient representation of your embedding vectors that supports real-time and approximate nearest neighbor (ANN) search queries. See Vector Search.
databricks vector-search-indexes create-index
Create a new index.
databricks vector-search-indexes create-index NAME ENDPOINT_NAME PRIMARY_KEY INDEX_TYPE [flags]
Arguments
NAME
Name of the index
ENDPOINT_NAME
Name of the endpoint to be used for serving the index
PRIMARY_KEY
Primary key of the index
INDEX_TYPE
There are two types of Vector Search indexes, Delta sync (an index that automatically syncs with a source Delta Table) or direct access (an index that supports direct read and write of vectors and metadata). Supported values: DELTA_SYNC, DIRECT_ACCESS
Options
--json JSON
The inline JSON string or the @path to the JSON file with the request body.
Examples
The following example creates a new Delta Sync index:
databricks vector-search-indexes create-index my-index my-endpoint id DELTA_SYNC
The following example creates a new Direct Access index:
databricks vector-search-indexes create-index my-index my-endpoint id DIRECT_ACCESS
databricks vector-search-indexes delete-data-vector-index
Delete data from index.
databricks vector-search-indexes delete-data-vector-index INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the vector index where data is to be deleted. Must be a Direct Vector Access Index.
Options
--json JSON
The inline JSON string or the @path to the JSON file with the request body.
Examples
The following example deletes data from a Direct Vector Access index:
databricks vector-search-indexes delete-data-vector-index my-index --json '{"primary_keys": ["key1", "key2"]}'
databricks vector-search-indexes delete-index
Delete an index.
databricks vector-search-indexes delete-index INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the index
Options
Examples
The following example deletes a vector search index:
databricks vector-search-indexes delete-index my-index
databricks vector-search-indexes get-index
Get an index.
databricks vector-search-indexes get-index INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the index
Options
Examples
The following example gets details for a vector search index:
databricks vector-search-indexes get-index my-index
databricks vector-search-indexes list-indexes
List indexes.
databricks vector-search-indexes list-indexes ENDPOINT_NAME [flags]
Arguments
ENDPOINT_NAME
Name of the endpoint
Options
--page-token string
Token for pagination
Examples
The following example lists all indexes in an endpoint:
databricks vector-search-indexes list-indexes my-endpoint
The following example lists indexes using pagination:
databricks vector-search-indexes list-indexes my-endpoint --page-token next-page-token
databricks vector-search-indexes query-index
Query an index.
databricks vector-search-indexes query-index INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the vector index to query
Options
--filters-json string
JSON string representing query filters
--json JSON
The inline JSON string or the @path to the JSON file with the request body.
--num-results int
Number of results to return
--query-text string
Query text
--query-type string
The query type to use
--score-threshold float
Threshold for the approximate nearest neighbor search
Examples
The following example queries an index with text:
databricks vector-search-indexes query-index my-index --query-text "search term" --num-results 10
The following example queries an index with filters:
databricks vector-search-indexes query-index my-index --query-text "search term" --filters-json '{"category": "product"}' --num-results 5
databricks vector-search-indexes query-next-page
Query next page.
databricks vector-search-indexes query-next-page INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the vector index to query
Options
--endpoint-name string
Name of the endpoint
--json JSON
The inline JSON string or the @path to the JSON file with the request body.
--page-token string
Page token returned from previous QueryVectorIndex or QueryVectorIndexNextPage API
Examples
The following example queries the next page of results:
databricks vector-search-indexes query-next-page my-index --endpoint-name my-endpoint --page-token "next-page-token"
databricks vector-search-indexes scan-index
Scan an index.
databricks vector-search-indexes scan-index INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the vector index to scan
Options
--json JSON
The inline JSON string or the @path to the JSON file with the request body.
--last-primary-key string
Primary key of the last entry returned in the previous scan
--num-results int
Number of results to return
Examples
The following example scans an index:
databricks vector-search-indexes scan-index my-index --num-results 100
The following example scans an index starting from a specific key:
databricks vector-search-indexes scan-index my-index --last-primary-key "key123" --num-results 100
databricks vector-search-indexes sync-index
Synchronize an index.
databricks vector-search-indexes sync-index INDEX_NAME [flags]
Arguments
INDEX_NAME
Name of the vector index to synchronize. Must be a Delta Sync Index.
Options
Examples
The following example triggers synchronization for a Delta Sync index:
databricks vector-search-indexes sync-index my-delta-sync-index
databricks vector-search-indexes upsert-data-vector-index
Upsert data into an index.
databricks vector-search-indexes upsert-data-vector-index INDEX_NAME INPUTS_JSON [flags]
Arguments
INDEX_NAME
Name of the vector index where data is to be upserted. Must be a Direct Vector Access Index.
INPUTS_JSON
JSON string representing the data to be upserted
Options
--json JSON
The inline JSON string or the @path to the JSON file with the request body.
Examples
The following example upserts data into a Direct Vector Access index:
databricks vector-search-indexes upsert-data-vector-index my-index '{"vectors": [{"id": "1", "vector": [0.1, 0.2, 0.3], "metadata": {"title": "Document 1"}}]}'
Global flags
--debug
Whether to enable debug logging.
-h or --help
Display help for the Databricks CLI or the related command group or the related command.
--log-file string
A string representing the file to write output logs to. If this flag is not specified then the default is to write output logs to stderr.
--log-format format
The log format type, text or json. The default value is text.
--log-level string
A string representing the log format level. If not specified then the log format level is disabled.
-o, --output type
The command output type, text or json. The default value is text.
-p, --profile string
The name of the profile in the ~/.databrickscfg file to use to run the command. If this flag is not specified then if it exists, the profile named DEFAULT is used.
--progress-format format
The format to display progress logs: default, append, inplace, or json
-t, --target string
If applicable, the bundle target to use