Forum Discussion
Table.Group sum up dynamic columns
I have a pivoted table where each column represents revenue data by week starts. Week starts are column headers. There could be various number of columns.
I need to sum up each column (total revenue by week). I was thinking to use Table.Group
(table as table, key as any, aggregatedColumns as list, optional groupKind as nullable number, optional comparer as nullable function) as table
aggregatedColumns should be something like this: {{"1/7/2015", each List.Sum([1/7/2015]), type number}....
But how can I generate it on fly for a dynamic list of columns, as I don't know how many columns do I have?
2 Replies
Got away without using Table.Group()... Unpivoted the table and grouped by dates to calculate sum of revenues per each date. Then pivoted the table by dates again...
You can handle this in Power Query (M) by dynamically building the list of aggregation operations. Since your week-start columns are dynamic, you can’t hardcode them — you generate the list from the column names.
Example in M:
let
Source = YourPivotedTable,
ColNames = List.RemoveItems(Table.ColumnNames(Source), {"KeyColumn"}), // keep out keys
AggList = List.Transform(ColNames, each {_, each List.Sum(_), type number}),
Result = Table.Group(Source, {"KeyColumn"}, AggList)
in
Result
Explanation:Table.ColumnNames(Source) → gets all column names.
List.Transform → turns each into an aggregation spec {ColumnName, each List.Sum([ColumnName]), type number}.
Table.Group → applies it dynamically.
This way, no matter how many week columns you have, the grouping/summing will work.
