Additional Indexes

Previous Next

See Also

USoft Developer will automatically generate indexes on all keys (primary, secondary, etc.) you defined for the various database objects. The naming convention for these generated indexes is:

<table_name>_I <key_number>

where 1 <= key_number <=9 (e.g. DEPT_I2).

To further optimize performance, you may define additional indexes on other (combinations of) fields. In particular, adding (additional) indexes to foreign key fields may greatly enhance the performance of your application. After defining indexes in the repository, you must have them actually created in the database.

If you define your own indexes, you are strongly advised NOT to use the above naming convention. Because the table generator looks up indexes in the databases using this naming convention, tables may be unnecessarily recreated.

NOTE:

If you have defined additional indexes on subtype tables that have NOT been implemented as separate tables, you MUST recreate the supertype tables involved.

To increase RDBMS performance, consider the following:

· Define an index on every foreign key.

 

· Make the selectivity of every index as great as possible.

 

· Define an index on fields that are often used in constraint specifications.

 

· Define an index on fields that are often used in queries (e.g. name, ZIP code, city, etc.)

Related Topics

How to Define Additional Indexes