> Hello all
>
> we developped a software with a data warehouse in sql server, without
using
> analysis services, and following the star design.
> we are working with data about transactions, with caracteristics modeled
in
> dimensions. in general each transaction has one single value for each
> caracteristic, but for one, it may have more than one or no value.
> some solution we thought about:
> -we use comma separated IDs (string) in the dimension in the fact table,
but
> it would hurt indexing
> -or store the IDs in a separate table, but each query will require a join
>
> my question is what is the best way to implement this feature ?
>
> thank you for your help