It is "Analysis Services:Choosing dimension types in sql server 2000
analysis services" I found it
under the msdn-sql server.
Also. There are several good labs available for Analysis Server if you have
at www.microsoft.com/sql . I think its on a download site there. These labs
have documentation as well.
DW consultant Sweden
> Thx Thomas,
> I think my problem is cause by the relationship between my fact table and
> reference table (just a distinct list of all DATEs).
> I am working around it by directly using the DATEs on the fact table, it
> seems to work now. But thanks anyway..
> BTW, I cannot find the white paper you mentioned, can you post the link,
> key work to search..
> > Well Ray. I think you need to carefully describe the problem here. Like
> > relational database
> > structures you need a relation between the dimension table and the fact
> > table. It is always a good
> > idea to avoid nulls in fact- and dimension tables. Analysis server
> > something called ragged
> > hierarchies, if you have a problem with nulls in the dimension tables.
> > is a whitepaper available
> > on MSDN about this. In Olap service you will have to use another
> > Nulls in the facttable (foreign keys or measures) is not a good idea.
> > Thomas Ivarsson
> > DW consultant
> > Sweden
> > > I found that I am missing some data measures from current records, it
> > turns
> > > out these records have <NULL> value on one of the columns. This column
> > > one of the dimension of my cube (SQLServer7). When the cube try to
> > > distinct values from this column to build the dimension, it can a 
> > >  month  day member, I think this is the way it interpret my <NULL>
> > > records, unforturately it don't know how to relation this back when it
> > build
> > > the aggregations.
> > > Any work around for this? What value is Time...? My only
> > > set "01/01/1900" to all <NULL>.
> > > Thank in advance