Forum Discussion
Performance hit after model change
The problem was actually the model, or maybe rather DAX generated while cross filtering. Power BI created one filter variable that contained filtering rules comming from two dimension tables around the fact table. Breaking it into two in DAX Studio made things work within eye blink but as I had no influence on code Power BI generates I simply changed the model, so one of the measure tables got joined into fact table and now it works fine. My measurements also appear to be good as report performance is now good enough for my users.
Hi,
thank you for your response. Your formula seems to speed things up significantly, as the run time of whole query dropped donw to like 45 seconds. What about the other one? Do you have any idea? Unfortunately I still didn't figured it out. The main problem I have there is no need of calculating much as this value is already assigned in every column, so I would be theoretically possible to just put it in the table but then when I summarize the table column in my visualization and two employees worked on the same task the value gets doubled and it is wrong but I still need column sum on the bottom. Can you help with that?
The problem was actually the model, or maybe rather DAX generated while cross filtering. Power BI created one filter variable that contained filtering rules comming from two dimension tables around the fact table. Breaking it into two in DAX Studio made things work within eye blink but as I had no influence on code Power BI generates I simply changed the model, so one of the measure tables got joined into fact table and now it works fine. My measurements also appear to be good as report performance is now good enough for my users.