As you can see from my example, the window is filled with just 2 filter options, because the 'persondetailsequenceno' contains so many different values.
But the thing is, this is a column where you will (almost) never filter on, but yet it takes almost the whole window.
Also i'm missing a search box at the top, which would make it already a bit user friendly.
Maybe you should be able somewhere to define if a column should popup in this window as an expanded filter or not ? That's probably not that easy to implement though.
I just tried rc2 and agree with this comment. I think the thing I'm missing most is the search box at the top. We've got some quite wide tables with lots of columns and relationships, this new filter window makes it really hard to find the right one to filter.
Listing all the category options by default makes this worse as you can only see 1 or 2 field names. This results in lots of scrolling and visually searching the screen.
The fields don't appear to be listed in alphabetical, database, or any other order (smart, custom from Metabase data model) that I can see.
It looks and works great in the sample data, but I'm struggling to see how this is an improvement on real world data sets. @flamber
I'm afraid for the 0.44.0 release, that nothing will change anymore.
Maybe later, but like you said, this looks nice on a simple database, like the demo one.
But not practicle on big databases with foreign keys and a lot of columns etc ..
Ok this is idd way better.
I do 'think' there is a little bug in it.
The column "Regtype" you see in the picture, is showing duplicate values.
1 = Registratie
2 = Saldo aanzuivering
3 = Saldo aanpassing
It is showing the 'real' database value AND the value that i have put in the data model (manually linking).
I already deleted and rescanned that field in the model, but with the same outcome.
Also, how are these filter ordend? It's not the same like the real data structure.
