A semantic_text field cannot be declared here. It is how a legacy
integrated index made by Indexes.createForModel reports its embedded
text field, so it still appears on the schema of an index you describe, as
one of the IndexSchemaField types. To have Pinecone embed text in an
index you create, declare a string field with embed or sparseEmbed.
Values you only need to filter on — numbers, booleans, string lists, and
plain strings — do not belong in the schema. Send them as document metadata
instead: they are indexed automatically at upsert time and appear on the
described index's IndexSchema.
The configuration of a single field in the schema of a new index.
A schema declares the searchable fields of the index. The creatable field types:
dense_vector— fixed-dimension vectors for semantic search.sparse_vector— sparse vectors for keyword or hybrid search.stringwithfullTextSearch,embed,sparseEmbed, or any mix of the three — see FullTextSearchStringField and IntegratedEmbeddingStringField.A
semantic_textfield cannot be declared here. It is how a legacy integrated index made by Indexes.createForModel reports its embedded text field, so it still appears on the schema of an index you describe, as one of the IndexSchemaField types. To have Pinecone embed text in an index you create, declare astringfield withembedorsparseEmbed.Values you only need to filter on — numbers, booleans, string lists, and plain strings — do not belong in the schema. Send them as document metadata instead: they are indexed automatically at upsert time and appear on the described index's IndexSchema.