The load process generates a unique ROWID_OBJECT value for the new record.

The load process performs foreign key lookups and substitutes any foreign key value(s) required to maintain referential integrity.

The load process inserts the record into the base object, and copies into this new record the generated ROWID_OBJECT value (as the primary key for this record in the base object), any foreign key lookup values, and all of the column data from the staging table (except PKEY_SRC_OBJECT)—including null values.

For each new record in the base object, the load process inserts the ROWID_OBJECT value into the associated dirty table so that match keys can be regenerated during the tokenization process.

For each new record in the base object, the load process sets its CONSOLIDATION_IND to 4 (ready for match) so that the new record can matched to other records in the base object.

The load process inserts a record into the cross-reference table associated with the base object. The load process generates a primary key value for the cross-reference table, then copies into this new record the generated key, an identifier for the source system, and the columns in the staging table (including PKEY_SRC_OBJECT).

If history is enabled for the base object, then the load process inserts a record into its history and cross-reference history tables.

If trust is enabled for one or more columns in the base object, then the load process also inserts records into control tables that support the trust algorithms, populating the elements of trust and validation rules for each trusted cell with the values used for trust calculations. This information can be used subsequently to calculate trust when needed.

If Generate Match Tokens on Load is enabled for a base object, then the tokenize process is automatically started after the load process completes.