Cognos Cube unicode

Cognos 8 Business Intelligence Transformer
1

Hi All!

I am new to Cognos8 . we have just started with Cognos. Our datawarehouse is based on a MS-SQL database. We have made the model in Framework and we are having problem with the transformer.

Our data contain Greek,Latin and Cyrillic characters. The cube is being biult correctly but when i try to access from the report or query or analysis the data that contain cyrillic i get an error and nothing is displayed.

Is there any workaround on this? It is a stop issue for us and i dont beleive that a tool like Cognos does not support having these 3 langs together in a cube…

The model in framework shows everything correct i n reports or queries…

Elias

2

Hi,

what is the exact error message you get when you open the cube with multiple charactersets?

3

Elias,

Transformer embeds a characterset identifier into the mdc file when building the cube. This is based on the locale defined in Configuration Manager/Cognos Shared/Locale. This is used by Cognos 8 to convert the cubes characterset to Unicode. If the embedded characterset identifier does not match the actual coding of the data or it does not match the OS (server) locale, mismatches may occur. The data in PowerPlay on the same machine as Transformer may still display fine as there is no chracterset conversion occurring.

Cubes built on an English O/S will be tagged as Windows 1252 encoded, so multi-byte cyrillic or other extended languages, will be displayed as if they were special Latin-1 characters. Hence you will see accented characters in the corrupted strings.

Try it and let us know if it is working well.

Hope this helps! :slight_smile:

4

The error we are getting is : “ANS-MES-0003 A server error occurred. Unable to complete the action.”

So what you are saying is to try different encodings or that it will work if we use English.

Either way today i plan to build some cubes with different locales to see if we will have any difference.

5

hmmm

so transformer can build a cube and then i would expect that the when the cube is correct, you also can open the cube without problems. I would expect then only strange characters, but that the cube works fine.

  1. this error message… did you get this error in Analysis studio?
  2. my feeling right now it that you transformer model has an issue. If you make a copy of the transformer model and reduce the number of dimensions and perhaps levels. Does this solve the ANS-MES-0003 message?
  3. how complex is this cube? ie. size, dimensions and levels and members?
6

hi all!

Guys the reply that i got from Cognos is :

“Content locale consists of the language and region combinations that are supported for the following:
· Content Manager objects
· Framework Manager objects
· Content Manager data formatting
· report data formatting
· report text (titles)
· database data, if the database tool is configured to use locale
Content locale does not apply to PowerCubes, whether they are created in IBM Cognos 8 Transformer or IBM Cognos Series 7 Transformer. Each PowerCube is created with a single language encoding. For multiple languages, a separate PowerCube is required for each language.”

As you can see there is no way to see Greek, english and Cyrillic on the same CUBE.

The reports that are based on the framework models wok fine…

So until they release a hotfix or something it is not possible to get analysis studio to show everything…If this is the case it is very strange that they allow this to happen.