Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!
I was working on a solution to calculate something similar to Qlikview's FirstSortedValue in DAX for Power BI. To my knowledge there is no solution out of the box to get the functionality of getting a text value (much like the category) which by its rank is in an nth position.
My question is: Why does this work on a single column, but returns a multiple column error when used on a table?
CALCULATE(SAMPLE(1, FILTER('table_name', RANKX(ALL('table_name'[column_name]), CALCULATE(COUNT('table_name'[column_name])),,,skip) = Nth), TRUE()))
a different variation as well:
CALCULATE(SAMPLE(1, VALUES('table_name'[column_name]), TRUE()), FILTER('table_name', RANKX(ALL('table_name'[column_name]), CALCULATE(COUNT('table_name'[column_name])),,,Dense) = Nth))
Either of these and many more of my attempts work when it is a single column, but why would this not work when used on a table with multiple columns? It should return a sample of 1 text value according to the Nth rank I choose? Akin to answering a question of "What is the Nth top selling product name according to salescount?".
I am stumped, still trying to find an answer. Firstnonblank is not an option, because I do not care for a solution to get the 1st ranked answer only; do not suggest that. 🙂
Other than that - please help, I feel I am close... 🙂
Solved! Go to Solution.
**EDIT: Oct 23 2018
Here it is! The final succinct version: works great!**
Measure =
VAR Nth = [number] RETURN CONCATENATEX(DISTINCT('table_name'[column_name]), CALCULATE(FILTER(VALUES('table_name'[column_name]), RANKX(ALL('table_name'[column_name]), CALCULATE(COUNT('table_name'[column_name])),,,Dense) = Nth)),",")
Hi @Anonymous,
I could use SAMPLE function in a calculated table without any error prompt. But I cannot understand what is the purpose of "= Nth" in above formula. Could you please provide more description with some sample data? And please show us your desired output.
Best regards,
Yuliana Gu
**EDIT: Oct 23 2018
Here it is! The final succinct version: works great!**
Measure =
VAR Nth = [number] RETURN CONCATENATEX(DISTINCT('table_name'[column_name]), CALCULATE(FILTER(VALUES('table_name'[column_name]), RANKX(ALL('table_name'[column_name]), CALCULATE(COUNT('table_name'[column_name])),,,Dense) = Nth)),",")
User | Count |
---|---|
139 | |
113 | |
103 | |
77 | |
63 |
User | Count |
---|---|
135 | |
126 | |
110 | |
70 | |
61 |