Research I did on Character Set conversions
We could put both DBs on AL32UTF8. Right now Cognos is AL32UTF8 and DW is WE8ISO8859P1.
UTF8 is a logical superset of WE8ISO8859P1, but not a binary superset. Meaning some characters may have different binary representation. For most characters (A-Z, 0-9, and most keyboard symbols) the representation will be the same single byte representation in both cases. But some special characters (ä, ö, ©, €) that store as a single byte in WE will store as multi-byte in UTF8. There could potentially be some conversion issues. I read an example where three dots (…) stored as 3 characters in WE may convert to an upside down “?” when importing into UTF8.
We may have no conversion issues, but we should check to be safe. Oracle provides a tool (cssscan) that will check for character conversion – we could run that to make sure everything would convert ok. It is a fairly common thing to convert character sets, so Oracle has the tools and capabilities to make it fairly easy. We just have to check this out properly.
Comments are currently closed.