I am in the process of changing the character set for our database. I have a couple questions with regard to the in-place migration using the CSALTER script and performing the full import/export. According to the Oracle documentation, the use of the CSALTER script requires that
it can be used only if all of the schema data is a strict subset of the new character set
Source:Changing the Database Character Set of an Existing Databasee
The current character set is WE8ISO8859P1. However, I want to change it to AL32UTF8. Is WE8ISO8859P1 a subset of AL32UTF8? And judging by experience, is it preferable to perform the migration using the CSALTER script as opposed to doing a full export/import?
Best Answer
Yes, WE8ISO8859P1 is a subset of AL32UTF8, though a small bit of conversion might be needed (which CSALTER will deal with & CSSCAN will inform you of when you do a preliminary scan).
A must-read for Oracle 10.2 is here.
Doing a full export/import will be more time-consuming than using csscan/csalter.
Another good, albeit old, read is the Oracle Character Migration Best Practices white paper.