Get certified for free when you join Fabric Data Days 2026 and dive into Fabric, Power BI, SQL, AI, and other essential data skills.
Join nowData Days is here! Join us now for 60+ days of learning, challenges, and connection. Learn more
Hello,
I need to know how to not count the extra rows power bi creates when I unpivot multiple columns?
Im working on survey data so all the questions I am unpivoting to get the responses for each question on the bottom tile.
the other tiles have filtering enabled and drillthrough. The pie chart is count by region and since I unpivoted all the questions it created alot of extra rows and the count of region is way off. Is there a way to not count those for that tile only?
Thanks!
Solved! Go to Solution.
Hi @Anonymous
I can reproduce your problem
T
To solve this, insert a step before "Unpivot columns"
add index columns from 1
then create a measure to calculate the count of regions
Measure = CALCULATE(DISTINCTCOUNT(Sheet5[Index]),ALLEXCEPT(Sheet5,Sheet5[region]))
Best Regards
Maggie
Hi @Anonymous
I can reproduce your problem
T
To solve this, insert a step before "Unpivot columns"
add index columns from 1
then create a measure to calculate the count of regions
Measure = CALCULATE(DISTINCTCOUNT(Sheet5[Index]),ALLEXCEPT(Sheet5,Sheet5[region]))
Best Regards
Maggie
Thank you so much for providing this solution. Works perfectly.
Hi @Anonymous
Could you show an example dataset after unpivoting columns?
Also, please change the pie chart and chart on bottom to table visual so i can see what data is in the visual.
Best Regards
Maggie
Hello @v-juanli-msft Thanks for the reply!
Here the example of the unpivoted dataset:
This has a total of 515 rows and after unpivoting it the row count gets to 7493. Total of 65 questions and their responses.
Dashboard data:
Here is a comparison on the count difference:
Let me know if you need anything else. Thanks again!
Sounds like you perhaps need a measure does does a COUNT but then divides that number by the COUNT of the DISTINCT (or VALUES) number of questions?
@Greg_Deckler Thanks for the quick response!
You are saying doing a count on the values (responses) and then divide by the count(distinct) of questions?
Thanks!
That was the idea, yes
@Greg_Deckler I tried:
Don't miss out on Data Days, June 15 through August 7. Learn Fabric, Power BI, SQL, AI and more.
Check out the May 2026 Power BI update to learn about new features.
| User | Count |
|---|---|
| 23 | |
| 23 | |
| 21 | |
| 20 | |
| 15 |
| User | Count |
|---|---|
| 58 | |
| 54 | |
| 42 | |
| 30 | |
| 24 |