Forum Discussion
Sort by value
Hi all,
I am trying to sort these table columns by values (column revenue), so the higher total value will be the first from the left and so on.
I clicked on the sort by column but nothing happened.
Thanks
Hi All,
I found a workaround solution.
After I changed between Rows and Columns parameter (ID will be Rows instead of Date) and apply column subtotal, I was able to sort by Sum of revenue values.
So the final result wasn't exactly like I wanted to but we will have to adjust it.
12 Replies
My screenshot is on Matrix as well.
Hi barlevitzky,
I'm not sure the Matrix visual can do what you're asking, but give this a shot:
In your matrix, click on the elipses at the top right. In the dialogue box that appears, see if there is a "sort by 'x'" section that will do what you're looking for.
Alright, I think I may have it. You'll want to create a new table that can handle the totaling for you, by doing the following:
1. create a new table that houses the values you're looking to use:
Table2 = SUMMARIZE( 'table' , 'table'[ID] )
2. calculate the totals in a new column that show up at the bottom of the Matrix:
Sum of Revenue= CALCULATE ( SUM( 'table'[revenue] ) , FILTER( 'table' , 'Table2'[ID] = 'table'[ID] ))
3. Create a rank column of these values:
Rank = RANKX( ALL( 'Table2' , 'Table2'[Sum of Revenue] , , DESC )
Then connect the 'Table2'[ID] column to your original data (the "one" side will be the new table), replace the columns in the Matrix visual with the new 'Table2'[ID], and Sort by Column on the Rank you just created.
EDIT: I realized that the Sort by Column feature won't let you sort based on the Rank, due to the FILTER being applied to get it. So as a workaround, you could concatenate the Rank and ID columns and use it as the column heading in your Matrix:
RankID = 'Table2'[Rank] & ". " & 'Table2'[ID]
Hi All,
I found a workaround solution.
After I changed between Rows and Columns parameter (ID will be Rows instead of Date) and apply column subtotal, I was able to sort by Sum of revenue values.
So the final result wasn't exactly like I wanted to but we will have to adjust it.
