Oracle | DBMS_VECTOR

Oracle DBMS_VECTOR_CHAIN Version 26ai

General Information

Library Note

Morgan's Library Page Header

The best Oracle News for FY2026

Oracle Database 26ai is now availble. If you haven't you downloaded already: Why?

Purpose Internal utilities supporting In Database Artificial Intelligence (AI) searches.

AUTHID CURRENT_USER

Constants

Name	Data Type	Value
Reference Constants
REFERENCE_PROFILE	VARCHAR2(7)	'PROFILE'
REFERENCE_CHUNKER	VARCHAR2(7)	'CHUNKER'
REFERENCE_EMBEDDER	VARCHAR2(8)	'EMBEDDER'
REFERENCE_INDEX	VARCHAR2(5)	'INDEX'
VECTORIZER	VARCHAR2(10)	'VECTORIZER'

Dependencies

ALL_CREDENTIALS	DRIG	JSON_KEY_LIST
DBMS_ASSERT	DRUYTK	JSON_OBJECT_T
DBMS_CRYPTO	DRIXMD	PLITBLM
DBMS_LOB	DRUE	PRVT_ORACHAIN
DBMS_PDB_LIB	DRVUTL	UTL_ENCODE
DBMS_SEARCH	DRVXMD	UTL_HTTP
DBMS_STANDARD	DRDEF	UTL_I18N
DBMS_VECTOR	JSON_ARRAY_T	UTL_RAW
DR$LIB	JSON_ELEMENT_T	VECTOR_ARRAY_T

Documented Yes

Exceptions

Error Code	Reason
ORA-20001	EXCP_CRED_NOT_EXIST (Credential Not Found)
ORA-20002	EXCP_PROVIDER_ERR (Provider Error)
ORA-20003	EXCP_GU_ERR

First Available 23ai

Security Model Owned by CTXSYS with EXECUTE granted to PUBLIC.

Source {ORACLE_HOME}/ctx/admin/dbmsvectorchain.pkh
{ORACLE_HOME}/cts/admin/dbmsvectorchain.plb

Subprograms

CREATE_CREDENTIAL	GET_PATH_ID	UTL_TO_CHUNKS
CREATE_LANG_DATA	GET_PATH_NAME	UTL_TO_EMBEDDING
CREATE_PREFERENCE	HYBRID_SEARCH	UTL_TO_EMBEDDINGS
CREATE_VOCABULARY	LIST_MODELS	UTL_TO_GENERATE_TEXT
DROP_CREDENTIAL	LIST_PROVIDERS	UTL_TO_RERANK
DROP_LANG_DATA	RERANK	UTL_TO_SUMMARY
DROP_PREFERENCE	SIMILAR	UTL_TO_TEXT
DROP_VOCABULARY

CREATE_CREDENTIAL (new 23ai)

Undocumented dbms_vector_chain.create_credential( credential_name IN VARCHAR2, params IN JSON);

TBD

CREATE_LANG_DATA (new 23ai)

Undocumented dbms_vector_chain.create_lang_data(params IN JSON);

TBD

CREATE_PREFERENCE (new 23ai)

Undocumented dbms_vector_chain.create_preference( pref_name IN VARCHAR2, pref_type IN VARCHAR2, params IN JSON DEFAULT NULL);

TBD

CREATE_VOCABULARY (new 23ai)

Undocumented dbms_vector_chain.create_vocabulary(params IN JSON);

TBD

DROP_CREDENTIAL (new 23ai)

Undocumented dbms_vector_chain.drop_credential(crednetial_name IN VARCHAR2);

TBD

DROP_LANG_DATA (new 23ai)

Undocumented dbms_vector_chain.drop_lang_data(pref_name IN VARCHAR2);

TBD

DROP_PREFERENCE (new 23ai)

Undocumented dbms_vector_chain.drop_preference(pref_name IN VARCHAR2);

TBD

DROP_VOCABULARY (new 23ai)

Undocumented dbms_vector_chain.drop_vocabulary(vocab_name IN VARCHAR2);

TBD

GET_PATH_ID (new 23ai)

Undocumented dbms_vector_chain.get_path_id( index_name IN VARCHAR2, path_name IN VARCHAR2) RETURN NUMBER;

TBD

GET_PATH_NAME (new 23ai)

Undocumented dbms_vector_chain.get_path_name( idxid IN NUMBER, path_id IN NUMBER) RETURN VARCHAR2;

TBD

HYBRID_SEARCH(new 23ai)

Undocumented dbms_vector_chain.hybrid_search(query_spec IN JSON) RETURN JSON;

TBD

LIST_MODELS (new 23ai)

Undocumented dbms_vector_chain.list_models(params IN JSON DEFAULT NULL) RETURN CLOB;

TBD

LIST_PROVIDERS (new 23ai)

Undocumented dbms_vector_chain.list_providers(parms IN JSON) RETURN CLOB;

TBD

RERANK (new 23ai)

Undocumented dbms_vector_chain.rerank( query IN CLOB, documents IN JSON, params IN JSON) RETURN JSON;

TBD

SIMILAR (new 23ai)

Undocumented dbms_vector_chain.similar( index_name IN VARCHAR2, vector_query IN CLOB, num_results IN NUMBER, text_query IN CLOB DEFAULT NULL, result_mode IN VARCHAR2 DEFAULT dbms_search.chunk_mode, score_function IN VARCHAR2 DEFAULT dbms_search.min_score) RETURN JSON;

TBD

UTL_TO_CHUNKS (new 23ai)

Undocumented

Overload 1 dbms_vector_chain.utl_to_chunks( data IN CLOB, params IN JSON DEFAULT NULL) RETURN sys.vector_array_t;

TBD

Overload 2 dbms_vector_chain.utl_to_chunks( data IN VARCHAR2, params IN JSON DEFAULT NULL) RETURN sys.vector_array_t;

TBD

UTL_TO_EMBEDDING (new 23ai)

Undocumented

Overload 1 dbms_vector_chain.utl_to_embedding( data IN CLOB, params IN JSON DEFAULT NULL) RETURN VECTOR;

TBD

Overload 2 dbms_vector_chain.utl_to_embedding( data IN BLOB, modality IN VARCHAR2 DEFAULT 'image', params IN JSON DEFAULT NULL) RETURN VECTOR;

TBD

UTL_TO_EMBEDDINGS (new 23ai)

Undocumented

Overload 1 dbms_vector_chain.utl_to_embeddings( data IN CLOB, params IN JSON DEFAULT NULL) RETURN sys.vector_array_t;

TBD

Overload 2 dbms_vector_chain.utl_to_embeddings( data IN sys.vector_array_t, params IN JSON DEFAULT NULL) RETURN sys.vector_array_t;

TBD

UTL_TO_GENERATE_TEXT (new 23ai)

Undocumented

Overload 1 dbms_vector_chain.utl_to_generate_text( data IN CLOB, params IN JSON DEFAULT NULL) RETURN CLOB;

TBD

Overload 2 dbms_vector_chain.generate_text( text_data IN CLOB, media_data IN BLOB, media_type IN VARCHAR2 DEFAULT 'image/jpeg', params IN JSON DEFAULT NULL) RETURN CLOB;

TBD

UTL_TO_RERANK (new 23ai)

Re-ranks documents based on the query relevance dbms_vector_chain.utl_to_rerank( query IN CLOB, documents IN JSON, params IN JSON DEFAULT NULL) RETURN JSON;

TBD

UTL_TO_SUMMARY (new 23ai)

Undocumented dbms_vector_chain.utl_to_summary( data IN CLOB, params IN JSON DEFAULT NULL) RETURN CLOB;

TBD

UTL_TO_TEXT (new 23ai)

Undocumented

Overload 1 dbms_vector_chain.utl_to_text( data IN CLOB, params IN JSON DEFAULT NULL) RETURN CLOB;

TBD

Overload 2 dbms_vector_chain.utl_to_text( data IN BLOB, params IN JSON DEFAULT NULL) RETURN CLOB;

TBD