-
Notifications
You must be signed in to change notification settings - Fork 25.8k
ES|QL: Add support for semantic_text in knn function #132066
Copy link
Copy link
Closed
Labels
:Search Relevance/SearchCatch all for Search RelevanceCatch all for Search Relevance>featureTeam:Search RelevanceMeta label for the Search Relevance team in ElasticsearchMeta label for the Search Relevance team in Elasticsearchpriority:normalA label for assessing bug priority to be used by ES engineersA label for assessing bug priority to be used by ES engineers
Metadata
Metadata
Assignees
Labels
:Search Relevance/SearchCatch all for Search RelevanceCatch all for Search Relevance>featureTeam:Search RelevanceMeta label for the Search Relevance team in ElasticsearchMeta label for the Search Relevance team in Elasticsearchpriority:normalA label for assessing bug priority to be used by ES engineersA label for assessing bug priority to be used by ES engineers
Type
Fields
Give feedbackNo fields configured for issues without a type.
tracked in #123043
In the query DSL, the knn query support semantic_text fields, not just dense_vector.
We did this as a way to give users more advanced control in how semantic_text fields are queried, since the knn query has options like
kandnum_candidates.We need to make sure we do the same in ES|QL and have the knn function accept semantic_text fields.
ES|QL sees semantic_text fields as just text, since it uses field caps to determine the field types.
This means that the knn function will need to support text fields as well.
In the case where the knn function is used with actual text fields (not semantic_text) we will only be able to raise an error at the shard level, which is fine as long as it's bubbled up correctly.