Forum Discussion
Post PatronMany to many crossfiltering none
MiKeZZa By linking multiple fact tables to the same Date Dimension, you can do exactly what you have described.
What you will end up is not really a star schema, but conceptually each FACT table is still a star schema. Its not really a snowflake schema either as you can't cascading outwards. I personally refer to it as a sandwich schema, because i put the Dimension tables on the outside (bread) and the fact tables in the centre (meat). I'm sure someone has a more technical name.
From here you can easily place non-related data on common axis and show results.
Post PatronHi Anonymous thank you for your answer. To be honoust; I'm really familair with things like starschema's and other DWH-principles. So your story is really common to me. But for Excel purposes we want to give people the ability to pick one date and see many, NOT RELATED, facts that were 'active' this day. It can be 0, 1 or more facts in 1 table on 1 day.
So this is not possible in a normal starschema.
MiKeZZa By linking multiple fact tables to the same Date Dimension, you can do exactly what you have described.
What you will end up is not really a star schema, but conceptually each FACT table is still a star schema. Its not really a snowflake schema either as you can't cascading outwards. I personally refer to it as a sandwich schema, because i put the Dimension tables on the outside (bread) and the fact tables in the centre (meat). I'm sure someone has a more technical name.
From here you can easily place non-related data on common axis and show results.
- MiKeZZa
Yes; you are alright. The solution is not in making a complex full join, but by just don't making a relation between the facts, but only between fact-dim. That's clear. Been there, done that, but haven't realised it when the current situation changed (were already was a full join).