5 Replies Latest reply: Aug 30, 2010 9:54 AM by sybrand_b RSS

    Character Set Question -WE8MSWIN1252 to AL32UTF8

    792829
      My source is in Linux RHEL 5.4 - Oracle 10.2.0.4
      character set: WE8MSWIN1252

      My target is in Linux RHEL 5.4 - Oracle 11.1.0.7
      character set: "AL32UTF8"

      Q1. We will be using Informatica PowerExchange CDC capture to pull data from source to Target. Will there be any issues with data since the character set is different?

      Q2. If we use DBLinks and pull data from source to target, will there be any issues since the character set is different.

      The target has not yet been set up. It is advisable to have the target DW characterset to WE8MSWIN1252? What is the difference between these two ?

      Thanks
        • 1. Re: Character Set Question -WE8MSWIN1252 to AL32UTF8
          sb92075
          What is the difference between these two ?
          AL32UTF8 can hold or represent more & different characters than WE8MSWIN1252
          • 2. Re: Character Set Question -WE8MSWIN1252 to AL32UTF8
            sybrand_b
            The difference between these two is the former is a single byte characterset, and the second one is a multibyte characterset.

            That said you are likely to run into issues, as a single byte character will be a multi-byte character in the target database.
            You need to make sure you only have varchar2(30 CHAR) columns (the default is BYTE) and set nls_length_semantics to CHAR.

            If you don't do that, you are going to experience unpleasant surprises.

            --------------
            Sybrand Bakker
            Senior Oracle DBA
            • 3. Re: Character Set Question -WE8MSWIN1252 to AL32UTF8
              792829
              Thank You!

              Having said that there might be issues, are we better off having the DW NLS_CHARACTERSET set to WE8MSWIN1252 as well. NLS_LANG for both is set to AMERICAN.

              The NLS_LENGTH_SEMANTICS is set to BYTE in Source and Target.

              Thanks
              Kasthuri
              • 4. Re: Character Set Question -WE8MSWIN1252 to AL32UTF8
                sybrand_b
                I would make sure both source and target database are either both multibyte or singlebyte
                or all length specifications are in CHAR instead of bytes, and nls_length_semantics has been set to char.
                Both scenarios will have impact on the current (incorrect) situation.
                You choose your poison.

                ------------
                Sybrand Bakker
                Senior Oracle DBA
                • 5. Re: Character Set Question -WE8MSWIN1252 to AL32UTF8
                  792829
                  Thank You. Ours is an internal ERP system DW. The Source systems use a windows application that displayed characters differently and thats the reason they changed their character set to WE8MSWIN1252 from AL32UTF8.
                  Fortunately the DW is not in place yet. We are in the phase of setting it up.
                  What would be the concerns in setting it up with the characterset as WE8MSWIN1252 to avoid the conversion issues?

                  Edited by: 789826 on Aug 31, 2010 1:26 PM