This topic contains 1 reply, has 2 voices, and was last updated by 
-
Posts
-
I was studying Rob’s outstanding cheat sheet, “Power Pivot and Power BI: How the DAX Engine Calculates Measures“.
In it Rob advises to avoid “Multi-Hop” look-ups, if possible. The advantage being better performance because a query isn’t hopping between tables, filtering them based upon relationship values, etc. The advice is to create a combined lookup table that joins the two more dimension tables.
Okey-dokey. So, here’s my question. With a combined lookup table you can loose the one-to-many relationships between the fact table and a lookup table. I believe a way around this would be to use the FILTER () function on the combined table based upon a value in the fact table. Is filtering a combined lookup table faster in performance than using multi-hop table relationships?
Anyone have an idea?
Ron
Ron,
You may be overcomplicating things. I have used snowflake schemas (multi-level dimension tables) and star schemas (single level dimension tables( in Power Pivot, and I have also converted snowflake schemas into star schemas like Rob recommends and I do not recall having to take the extra filtering step you are considering.
The forum ‘Power Pivot’ is closed to new topics and replies.