NLS character encoding

If NLS character encoding is enabled on your source computer, you should enable it on your target computer. NLS settings should be the same on both computers. If NLS support is not installed on your source computer, then you need to consider the InfoSphere® DataStage® jobs that you are migrating, and the NLS settings that you want to use on your target computer.

You should enable NLS if you have data from many geographies that might include international symbols like the euro sign (€). InfoSphere Information Server cannot process some characters if NLS is not enabled. InfoSphere DataStage parallel jobs can migrate to an installation of InfoSphere Information Server with NLS enabled without any issues. InfoSphere DataStage server jobs are more difficult to migrate to NLS enabled computers. If you primarily have server jobs, you should consider the impact of migrating to an NLS enabled computer. The performance of server jobs is impacted and jobs with locale sensitive functions like ICONV and OCONV must be changed to include the locale information.

InfoSphere DataStage server jobs and parallel jobs that use server shared containers or the BASIC transformer stage are the jobs that are impacted by NLS settings. The following behavior occurs when you run these jobs with NLS enabled:

Server jobs are run with UTF-8 encoding instead of running in host character set bytes, which impacts performance.
Jobs that process non-ASCII characters might not work the same way. This includes extended ASCII, non-character byte streams and EBCDIC characters. Specifically, string handling functions with NLS encoding enabled processes data as UTF-8 characters that are 1 to 3 bytes in length.
Jobs with byte-based offsets that do not have NLS enabled will not work the same way as they did when NLS was enabled, unless they are only processing 7-bit ASCII data. This applies to all character based functions, including string indexing constructs of the form string[start,count].
Locale behavior changes from the InfoSphere DataStage defaults to locale-specific functions. This includes date and time representations and conversions, alphabetic sort orders, numeric representation, currency, and character classifications, including case handling.

The only time you can enable NLS character encoding for InfoSphere Information Server is during installation. If you install InfoSphere Information Server on your target computer without enabling NLS, then you cannot modify it post-installation.

If the language of your operating system is set to something other than English, the installation automatically enables NLS. If your operating system language is English, you are given the option to turn off NLS enablement during the installation process.

If you had NLS enabled on your source computer, you should check the default character set. If you are migrating from InfoSphere Information Server, Version 8.0.1 or earlier, or if you are changing platforms, your new computer might have a new default character set. If the default character set changes, you might need to change the default character set for the InfoSphere DataStage project or jobs if you do not want the behavior of the jobs to change. In addition, since sorts can be done at the operating system level, a character set change could result in small order changes. If you want to use the same character set on the target computer that you use on the source computer, then you need to set the character set on your operating system before installing prior to installing InfoSphere Information Server. For additional information, see Setting the locale and character encoding for the engine tier target computer.