Forum Discussion
Update calculated column values with slicer
Hey all,
I know that this topic has been addressed several times (computed columns do not response on slicer because they are computed upfront) but all workarounds did not work for my use case.
I am trying to compute a walkforward accuracy (1=accurate prediction; 0 = false prediction) in DAX that also depends on the selected date range.
Simplified example - (name of the table: Accuracy)
| Date | Validation | |
| 01.01.2020 | 0 | |
| 02.01.2020 | 1 | |
| 03.01.2020 | 1 |
When I am now slicing the date range to (02.01.2020 to 03.01.2020), I always get the blue line instead of the orange one (orange would be correct for my example). In several threads I read about modeling it as a measure but I failed to formulate it in DAX. Eventually, I want to show the walkforward accuracy (depending on date range) as a simple line diagram.
Not sure if that is of relevance but I am slicing the data based on another date dimension table that is in relationship (1:many) with Accuracy[Date].
Help is much appreciated!
2 Replies
nikextens Definitely needs to be a measure. But, I don't understand the expected calculation, where the 100% are supposed to come from. You may need a disconnected calendar table but hard to know.
Greg_Deckler For 02.01.2020 the signal is correct (Validation = 1) so the walkforward accuracy starts with 100% (1 out of 1 predictions were correct) and for the 03.01.2020 the signal is also correct (Validation = 1, thus, 2 out of 2 predictions are correct). Walkforward accuracy remains at 100%.
