A couple powerBI notes
I’ve been working on a big powerBI dog and pony show lately, and have discovered a couple things in my haste. (Not really when I want to be discovering them)
- Filtering over a summarize or crossjoin is a terribly inefficient way to write DAX. It will iterate over the result set, rather than apply the filter in context.
- Whiskey does not help writing DAX. It’s just like using the filter operator. You think you’re efficient, but you’re really not.
- A dimension will be restricted by the “fact” or measure tables it’s associated with if the relationship is set to “Both”. As this is the default, it’s easy to get screwed by this, especially if you’ve had several whiskey. In other words, if you’re associating a dimension with a fact, (or base table), it will look like the dimension has fewer keys than it actually does.