- The column you specify for the NORMALIZE option must have either a Period data type or be a derived Period column.
- If you do not specify an explicit normalization condition, the default is ON MEETS OR OVERLAPS.
- If a table that is to be normalized contains columns with a BLOB, CLOB, JSON, or XML data type, those columns must be explicitly specified in the normalize_ignore_column_name list for the NORMALIZE option. The XML data type is only supported on the Block File System on the primary cluster. It is not available for the Object File System.
If a table definition contains such a column that is not included in the normalize_ignore_column_name list, the system returns an error to the requestor.
- Vantage validates both UNIQUE and PRIMARY KEY constraints with normalized rows. If a normalized row violates a UNIQUE or PRIMARY KEY constraint, the system returns an error to the requestor.
- Vantage validates CHECK constraints on a row inserted into a normalized table, and if the constraint is violated, the system returns an error to the requestor.
This action prevents a security issue that can occur if a constraint is specified on the beginning or end of a normalized Period column. The input row violates the CHECK constraint but the normalized row does not. This situation cannot occur with UNIQUE constraints.