4 Replies Latest reply: Feb 17, 2012 4:27 PM by JustinCave RSS

    Migration Characterset Issue

    885137
      Hi,

      I am involved in a Data Migration project.
      The database version I am using is Oracle 10.2.0.5 and the NLS_CHARACTERSET is set to UTF8 on the database.
      The requirement is to provide the data extract in Unicode for the new databse.

      I have set the NLS_LANG variable to AL32UTF8 on my machine.
      For extracting the data from all the tables except one, I did not face any problems.
      But for one table,while exporting the data I got the error ORA:29275 and ORA-12703 while trying to convert the strings.
      I was able to address this issue by changing the characterset on my machine to UTF8.

      My Question is:

      1)Will the extract files pulled out in AL32UTF8 characterset pose a problem to the new database(if in UTF8) where the data has to be migrated?

      2)I got to know that Oracle UTF8 is not exactly "Unicode" as per the standards, so how do I address the dataset in UTF8 if to be provided in Unicode?

      thanks in advance
      deb
        • 1. Re: Migration Characterset Issue
          19426
          Migration Characterset Issue
          Globalization forum issue:

          Globalization Support
          • 2. Re: Migration Characterset Issue
            orafad
            It would seem this post should have been posted in {forum:id=50}.


            1. How exactly are you "extracting the data"?

            In general, when using export tool, you should use a client char set that matches the database, as to avoid or delay conversion.
            (Oracle character sets UTF8 and AL32UTF8 are not the same.)

            But ORA-29275 possibly indicates some problem with source data, so it is best to investigate the issue.

            Why does the new database have UTF8 over the current AL32UTF8? The first, UTF8 should be used only when (the very old) client applications require it.


            2. Where did you read this?

            See Globalization Support guide about Unicode versions and Oracle character sets.
            • 3. Re: Migration Characterset Issue
              885137
              Hi orafaq

              please let me know

              Will the extract files pulled out in AL32UTF8 characterset pose a problem to the new database(if in UTF8) where the data has to be migrated
              • 4. Re: Migration Characterset Issue
                JustinCave
                What, exactly, do you mean by "extract files"? What tool are you using to generate these files? Can you describe the process (including the NLS_LANG on any machines where it is relevant)?

                Do all the characters in the existing database exist in the earlier Unicode standard that a database character set of UTF8 implies? Or are you using some characters that were added in newer versions of the Unicode standard?

                Justin