Table Procedure Examples, Design, and Use › Choose Unique Key Fields › INSERT Operations

INSERT Operations

You must define unique keys to perform INSERT processing because only one row will be added to the table and all the owners in the primary path must be identified before the row can be inserted. The target record does not, however, require a unique key.

When a unique key is defined for the target record, the rules for the target record are the same as other path records in regard to "Next Row" processing. But CA IDMS does not determine if the target record is unique unless duplicates are not allowed for the record or set relationship. Therefore, the table procedure returns an IDMS error (1205), not an SQL error. Unique keys are also important to INSERT processing when the target record is inserted in an auxiliary set relationship with an Automatic connect option. Auxiliary records must have unique keys defined for them whenever the relationship is Automatic. Otherwise, IDMS currency is not established for that set, and the INSERT fails.