Integrating multilingual data

[sandy2008]

We need to integrate various source information from SQL, Oracle, Db2, Sybase, MYSQL to SQL Server 2005. Source data is in multilingual format. What could be a good approach to integrate multilingual data and how can this be achived?

 

[hwkranger]

This would be the same as any data warehousing development process.

Very carefully.

But seriously, for multi-linguage data, I expect that you have data keyed by language, I also expect that you should be building your dimension tables in respect to this.

Get a good inventory on your data, map it correctly, with solid due-diligence you'll do fine.

 

[Hans63]

One approach I can come up with is to find all textual descriptions (where the multilingual aspect comes out) and replace them by a text identifier. You then store all texts in a separate table, with this identifier and a language key.
It requires a thorough analysis of the source systems, to identify which descriptors are actually the same in various languages. And you need all this metadata to make your transformation work.
This is the approach which is used by several vendors of multilingual software and I think it should work here as well.

Another possible approach is to store all descriptive texts in separate fields in a table, but I really doubt this will work.