Choosing Character Sets and Collation Sequences to Maximize Performance

Performance for character data is better with a binary character set and collation sequence than with a nonbinary character set and collation sequence.

To maximize performance, create a database with these default option settings, which result in a binary character set and collation sequence:

CREATE DATABASE dbname
COLLATION 'ISO_BINENG' 
CASE RESPECT

Other settings for these options result in a nonbinary character set and collation sequence.

The disadvantage of these settings is that uppercase characters are always sorted before lowercase ones. For example, “BANANA” sorts before “apple.” If you prefer a more natural sort order, but still need a case-sensitive database, and if you are willing to sacrifice some degree of performance, use the collation ISO_1 instead of the default, ISO_BINENG.

Note: Passwords in newly-created databases are case-sensitive, regardless of the case-sensitivity of the database. When you rebuild an existing database, the password follows the case-sensitivity of the original database.