Siebel Global Deployment Guide > Overview of Global Deployments >
About the Database Collation Sequence
The collation sequence, also called sort order, is the ordering relationship, or sequence, between data records. Each database has a collation sequence so that records returned by queries can be returned in a certain order, such as an alphabetic order for text strings. The collation sequence determines the order in which records are displayed in the Siebel client, most noticeably in list views.
A collation sequence is defined when you set up the Siebel Database. All sorting is done in the Siebel Database by the database server. Sorting is not set or performed within the Siebel application and does not depend on the operating system.
NOTE: For more information about creating and configuring the Siebel Database, see the Siebel Installation Guide for the operating system you are using. For more information about collation sequences for upgrade environments, see Siebel Database Upgrade Guide.
For the collation sequences supported for each supported RDMBS platform for the Siebel Database, see Siebel System Requirements and Supported Platforms on Siebel SupportWeb. Also consult your RDBMS vendor documentation.
The collation sequence in effect for a database is determined by one of the following implementation methods:
- Indexes created in the Siebel Database provide a default collation sequence. In Oracle databases, indexes always use binary collation sequence.
- Post-query sorting may also be supported for an RDBMS platform. However, this method of sorting yields slower performance and requires all records to have been retrieved first. For this reason, it is impractical for Siebel applications, which always perform open-ended queries.
For the development environment, only binary collation sequence is supported. For a production environment, you can specify the collation sequence most suitable for your deployment.
NOTE: Changing the collation sequence after the Siebel Database has been installed requires rebuilding your indexes. On a fully loaded production database, this task is time-consuming and database resource-intensive. It is advisable to consult with Expert Services when planning a project of this complexity.
Which collation sequence is best for your deployment depends on factors such as RDBMS support, performance requirements, database availability requirements, the code page in use, the needs of your users, and the nature of the data that is to be retrieved by different groups of users.
Binary collation sequence offers the best performance and does not require you to rebuild your indexes for the production environment. This collation sequence works well for users working with English-language data, because the ASCII character set is based on the English alphabet and corresponds to the binary collation sequence. However, sorting may be unsuitable for users and data in languages other than English.
For multilingual deployments using Unicode, a linguistic collation sequence based on the Unicode Collation Algorithm (UCA), which goes by different names for different RDBMS vendors, may be a suitable collation sequence. UCA, also known as ISO 14651, provides reasonably good results with mixed-language data.
Other linguistic, or dictionary, collation sequences may offer optimal sorting results for particular languages or groups of languages. Such collation sequences may be suitable for certain deployments, such as those requiring compatibility with the CP932 (Japanese Shift-JIS) sort order.
Linguistic collation sequences that are not based on UCA may associate multiple characters (such as accented and unaccented versions of a particular letter) so they will be treated the same for sorting purposes, but will also be treated the same in unique indexes. If you are changing to a case- or accent-insensitive collation sequence, you will need to first clean out any data that is unique only due to a case or accent difference.
Database Collation for the Mobile Web Client
For the Siebel Mobile Web Client, which uses a local SQL Anywhere database, the default collation sequence for local database indexes is determined by the setting for the local database template from which each individual mobile user's local database is initialized.
Binary, UCA, and other linguistic collation sequences are available for the local database. Binary collation sequence provides the best performance, but other collation sequences may perform acceptably.
NOTE: Synchronization conflicts could occur when collation sequences with different case or accent sensitivity are in effect on the local database and the enterprise database. For this reason, choose a collation sequence on the local database that is at least as case or accent sensitive as the collation sequence on the enterprise database. Avoid choosing a case-insensitive collation sequence on the local database if your server database is case sensitive.
It is possible to specify a post-query collation sequence for an individual mobile user's local database, by specifying a value for the SORTCOLLATION parameter in the application configuration file. Performance issues may arise from specifying a post-query collation sequence in this manner.
For more information about the Mobile Web Client and the SORTCOLLATION parameter, see Siebel System Administration Guide, Siebel Remote and Replication Manager Administration Guide, and other applicable documentation on Siebel Bookshelf.
NOTE: A local database used for development with Siebel Tools must use the binary collation sequence. Using Siebel Tools against a non-binary collation sequence is not supported.