vector_search_indexes

Creates, updates, deletes, gets or lists a vector_search_indexes resource.

Overview

Name	vector_search_indexes
Type	Resource
Id	databricks_workspace.vectorsearch.vector_search_indexes

Fields

The following fields are returned by SELECT queries:

vector_search_indexes_get_auto_eval_status

Name	Datatype	Description
job_id	string	Databricks Jobs job_id of the autoeval background-compute job for this index, so the UI can surface a link to the job. Unset when no autoeval job exists for the index yet.
end_time_ms	integer
▶latest_run	object	State of the latest autoeval run, including stage progress. Populated only while status is AUTO_EVAL_DISPLAY_STATUS_RUNNING and the running wheel has reported at least one stage update. Absent for terminal states.
run_as_user	string	The user the latest job run was created as. Used by the UI to construct the per-run MLflow dashboard URL.
state_message	string	Free-form failure copy from the underlying job. Populated only when status is AUTO_EVAL_DISPLAY_STATUS_FAILED. Capped server-side to bound payload size when the job emits long stack traces.
status	string	Current display status of the latest autoeval run. (AUTO_EVAL_DISPLAY_STATUS_FAILED, AUTO_EVAL_DISPLAY_STATUS_PENDING, AUTO_EVAL_DISPLAY_STATUS_RUNNING, AUTO_EVAL_DISPLAY_STATUS_SUCCEEDED)

Methods

The following methods are available for this resource:

Name	Accessible by	Required Params	Optional Params	Description
vector_search_indexes_get_auto_eval_status	select	name, deployment_name		Returns the status of the latest autoeval run for a vector index.
run_auto_eval	exec	name, deployment_name		Triggers an autoeval quality evaluation for a vector index.
run_reranker_finetuning	exec	name, deployment_name		Triggers reranker finetuning for a vector index.
update_index_budget_policy	exec	index_name, deployment_name		Update the budget policy of an index

Parameters

Parameters can be passed in the WHERE clause of a query. Check the Methods section to see which parameters are required or optional for each operation.

Name	Datatype	Description
deployment_name	string	The Databricks Workspace Deployment Name (default: dbc-abcd0123-a1bc)
index_name	string	Name of the AI Search index
name	string	Fully qualified index name (catalog.schema.index).

SELECT examples

vector_search_indexes_get_auto_eval_status

Returns the status of the latest autoeval run for a vector index.

SELECT
job_id,
end_time_ms,
latest_run,
run_as_user,
state_message,
status
FROM databricks_workspace.vectorsearch.vector_search_indexes
WHERE name = '{{ name }}' -- required
AND deployment_name = '{{ deployment_name }}' -- required
;

Lifecycle Methods

run_auto_eval

Triggers an autoeval quality evaluation for a vector index.

EXEC databricks_workspace.vectorsearch.vector_search_indexes.run_auto_eval 
@name='{{ name }}' --required, 
@deployment_name='{{ deployment_name }}' --required 
@@json=
'{
"num_queries": {{ num_queries }}, 
"num_results": {{ num_results }}, 
"query_types": "{{ query_types }}", 
"queryset_query_column": "{{ queryset_query_column }}", 
"queryset_relevant_docs_column": "{{ queryset_relevant_docs_column }}", 
"queryset_table": "{{ queryset_table }}"
}'
;

run_reranker_finetuning

Triggers reranker finetuning for a vector index.

EXEC databricks_workspace.vectorsearch.vector_search_indexes.run_reranker_finetuning 
@name='{{ name }}' --required, 
@deployment_name='{{ deployment_name }}' --required 
@@json=
'{
"embedding_model": "{{ embedding_model }}", 
"model_name": "{{ model_name }}", 
"num_queries": {{ num_queries }}, 
"query_column": "{{ query_column }}", 
"query_table": "{{ query_table }}"
}'
;

update_index_budget_policy

Update the budget policy of an index

EXEC databricks_workspace.vectorsearch.vector_search_indexes.update_index_budget_policy 
@index_name='{{ index_name }}' --required, 
@deployment_name='{{ deployment_name }}' --required 
@@json=
'{
"usage_policy_id": "{{ usage_policy_id }}"
}'
;