When you are deploying a database worldwide, you should consider the locales in which the database is to be used. Different locales may have different sort orders or text comparison rules. For example, the database that you are going to deploy maybe have been created using the 1252LATIN1 collation and this may not be suitable for some of the environments in which it is to be used.

Since the collation of a database cannot be changed after it has been created, you might consider creating the database during the install phase and then populating the database with the schema and data that you want it to have. The database can be created during the install either by using the dbinit utility, or by starting the database server with the utility database and issuing a CREATE DATABASE statement. Then, you can use SQL statements to create the schema and do whatever else is necessary to set up your initial database.

If you decide to use the UCA collation, you can specify additional collation tailoring options for finer control over the sorting and comparing of characters using the dbinit utility or the CREATE DATABASE statement. These options take the form of keyword=value pairs, assembled in parentheses, following the collation name. Using the CREATE DATABASE statement, for example, a collation tailoring can be specified using syntax like the following:

CHAR COLLATION 'UCA( locale=es;case=respect;accent=respect )'

As an alternative, you could create multiple database templates, one for each of the locales in which the database is to be used. This may be suitable if the set of locales you are deploying the database to is relatively small. You can have the installer choose which database to install.

See also

• CREATE DATABASE statement
• Initialization utility (dbinit)
• Choosing collations