This illustration shows that creating a ubiquitous search index mainly involves two commands. First, the DBMS_SEARCH.CREATE_INDEX API that creates an index table named MYINDEX. This index table contains the DATA, METADATA, OWNER, SOURCE, and KEY columns. A MYINDEX index is also shown as created on the DATA column. Here, the schema owner name SCOTT is also specified along with the index name as SCOTT.MYINDEX.

The second command shown is the DBMS_SEARCH.ADD_SOURCE API to add the PRODUCTS and CUSTOMERS table into MYINDEX. Here too, the schema owner name is specified along with the source table name as SCOTT.PRODUCTS and SCOTT.CUSTOMERS.

The DATA column is shown as an empty, and it is indicated that data from both the PRODUCTS and CUSTOMERS tables is being added to this index.

The METADATA column contains entries for each row of both the source tables as:

{"OWNER":"SCOTT","SOURCE":"PRODUCTS","KEY":{"ID":1}}
{"OWNER":"SCOTT","SOURCE":"PRODUCTS","KEY":{"ID":2}}
{"OWNER":"SCOTT","SOURCE":"CUSTOMERS","KEY":{"ID":5}}
{"OWNER":"SCOTT","SOURCE":"CUSTOMERS","KEY":{"ID":9}}