The rules for the MODIFY options for non-character partitioning in General Rules for the MODIFY PRIMARY Clause and Rules for the MODIFY For Multilevel Partitioning also apply to character partitioning and to the character row partitioning levels of a column-partitioned table or join index.
The following rules are specific to character row partitioning levels:
- You can specify a partitioning expression that involves character or graphic comparisons such as =, >, ≤, BETWEEN, or LIKE.
- A character row partitioning expression cannot specify Kanji1 or KanjiSJIS columns or constant expressions.
- A character row partitioning expression must be based on one or more columns from the table or deterministic expressions based on the columns from the table or join index the partitioning expression is written for.
- The expressions and referenced columns in the specification of the character partitioning expression must not have CLOB data types.
- The expressions specified in a character row partitioning expression must not contain any of the following functions.
- Aggregate functions
- Built-in functions
- Grouped row OLAP functions
- UDFs of any kind
- HASHAMP function
- HASHBAKAMP function
- RANDOM function
- The expressions specified in a row partitioning expression must not contain any of the following system-derived columns, operators, or SELECT subqueries.
- ROWID
- PARTITION
- PARTITION#L n
- Set operators
- Subqueries