How LOAD orders loaded records
The LOAD utility loads records into a table space in the order in which they appear in the input stream. It does not sort the input stream, and it does not insert records in sequence with existing records, even if a clustering index exists.
To achieve clustering when loading an empty table or replacing data, sort the input stream. When adding data to a clustered table, consider reorganizing the table after running LOAD.
Because rows with duplicate key values for unique indexes fail to be loaded, any records that are dependent on such rows either:
- Fail to be loaded because they would cause referential integrity violations (if you specify ENFORCE CONSTRAINTS)
- Are loaded without regard to referential integrity violations (if you specify ENFORCE NO)
As a result, violations of referential integrity might occur. Such violations can be detected by LOAD (without the ENFORCE(NO) option) or by CHECK DATA.