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 nowTry your skills in the Power BI Dataviz World Championship! Round one ends June 26. Join now
I need to find out how to calculate the average of a column of a table stored as a variable. I have data stored in the same format as A-C in the example below.
I want to find out how to average the top 3 values of the 5 most recent dates for each group. I think I have figured out how to get those top 3 values alone in a table stored as a variable, however I can't seem to average them.
I first created a table with a unique list of the groups:
Averages = distinct(Data[Group])
Then I tried to make a calculated column as follows:
Avgs = var GroupTable = filter(Data, Data[Group] = Averages[Group]) var DatesRankedTable = addcolumns(GroupTable, "DateRank", rankx(GroupTable, [Date])) var Top5Dates = filter(DatesRankedTable, [DateRank] <= 5) var ValueRankedTable = addcolumns(Top5Dates, "ValueRank", rankx(Top5Dates, [Value])) var Top3Values = filter(ValueRankedTable, [ValueRank] <= 3) return Average(Top3Values[Value])
If I do all the variable tables manually step by step, the "Top3Values" table comes out as it should. However then when I try to do the average of [Value] in "Top3Values" it errors out. How can I get the average of a column that is in this table stored as a variable.
Solved! Go to Solution.
I'm not sure why I can't average a column from a variable table. However I was able to get the result I wanted by using:
Avgs = var GroupTable = filter(Data, Data[Group] = Averages[Group]) var DatesRankedTable = addcolumns(GroupTable, "DateRank", rankx(GroupTable, [Date])) var Top5Dates = filter(DatesRankedTable, [DateRank] <= 5) var ValueRankedTable = addcolumns(Top5Dates, "ValueRank", rankx(Top5Dates, [Value])) var Top3Values = filter(ValueRankedTable, [ValueRank] <= 3) var AverageValue = sumx(Top3Values[Value])/countrows(Top3Values) return AverageValue
I'm not sure why I can't average a column from a variable table. However I was able to get the result I wanted by using:
Avgs = var GroupTable = filter(Data, Data[Group] = Averages[Group]) var DatesRankedTable = addcolumns(GroupTable, "DateRank", rankx(GroupTable, [Date])) var Top5Dates = filter(DatesRankedTable, [DateRank] <= 5) var ValueRankedTable = addcolumns(Top5Dates, "ValueRank", rankx(Top5Dates, [Value])) var Top3Values = filter(ValueRankedTable, [ValueRank] <= 3) var AverageValue = sumx(Top3Values[Value])/countrows(Top3Values) return AverageValue
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 |
|---|---|
| 24 | |
| 21 | |
| 20 | |
| 17 | |
| 12 |
| User | Count |
|---|---|
| 64 | |
| 55 | |
| 42 | |
| 38 | |
| 30 |