Forum Discussion
Helper ICalculate average based on dynamic column reference
Hello BartVW
in order to create a dynamically solution, you have to tell us what can be a rule, to select the correct columns. Here an example where I calculated the average of all columns that contain a ".". If this is your requirement, you can use the code, otherwise tell us all names of your columns, and we will find another logic. Replace the source-step and changedtype with your real data table
let Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUTICYmMgNgFiUyA2A2JzpdhYAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Column1 = _t, Column2 = _t, #"007.2020" = _t, #"008.2020" = _t, #"009.2020" = _t, #"010.2020" = _t, Result = _t]), #"Changed Type" = Table.TransformColumnTypes(Source,{{"Column1", Int64.Type}, {"Column2", Int64.Type}, {"007.2020", Int64.Type}, {"008.2020", Int64.Type}, {"009.2020", Int64.Type}, {"010.2020", Int64.Type}, {"Result", Int64.Type}}), AddAverageOfMonths = Table.AddColumn ( #"Changed Type", "AverageMonth", (row)=> List.Average(Record.FieldValues(Record.SelectFields(row, List.Select(Table.ColumnNames(#"Changed Type"), each Text.Contains(_,"."))))) ) in AddAverageOfMonthsCopy paste this code to the advanced editor in a new blank query to see how the solution works.
If this post helps or solves your problem, please mark it as solution (to help other users find useful content and to acknowledge the work of users that helped you)
Kudoes are nice too
Have fun
Jimmy
Community ChampionHello BartVW
in order to create a dynamically solution, you have to tell us what can be a rule, to select the correct columns. Here an example where I calculated the average of all columns that contain a ".". If this is your requirement, you can use the code, otherwise tell us all names of your columns, and we will find another logic. Replace the source-step and changedtype with your real data table
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUTICYmMgNgFiUyA2A2JzpdhYAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Column1 = _t, Column2 = _t, #"007.2020" = _t, #"008.2020" = _t, #"009.2020" = _t, #"010.2020" = _t, Result = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Column1", Int64.Type}, {"Column2", Int64.Type}, {"007.2020", Int64.Type}, {"008.2020", Int64.Type}, {"009.2020", Int64.Type}, {"010.2020", Int64.Type}, {"Result", Int64.Type}}),
AddAverageOfMonths = Table.AddColumn
(
#"Changed Type",
"AverageMonth",
(row)=> List.Average(Record.FieldValues(Record.SelectFields(row, List.Select(Table.ColumnNames(#"Changed Type"), each Text.Contains(_,".")))))
)
in
AddAverageOfMonths
Copy paste this code to the advanced editor in a new blank query to see how the solution works.
If this post helps or solves your problem, please mark it as solution (to help other users find useful content and to acknowledge the work of users that helped you)
Kudoes are nice too
Have fun
Jimmy
Helper IJimmy801 thank you, this works great and solves my immediate challenge.
As a PQ beginner, i would love to understand more how this works though, and had some questions:
- what is the "(row) ==> " syntax about? It looks like a function but without 'let' and 'in'?
- what is the reasoning to go from ColumnNames to List.Select to Record.SelectFields and Record.FieldValues? Why did my original solution not work, just using "each List.Average(List.FindText(Table.ColumnNames(#"Pivoted Column"),"202"))"
- why do you use the _ variable?
Appreciate all the help you can provide, and if you have any good links i could look at for these questions, that would be great too.
Cheers
Bart
- Jimmy801
Hello BartVW
If my answer solved the problem, I would appreciate if you could mark it as solution.
I will shortly try to answer your question, but you could search the documentation for this as well.
- ()=> is the syntax of a function (or each would also work). I could have used also let and in to split my steps, but i prefered to put everything in on line, so no let and in is needed
- columnnames gibs you the headernames, list select filters according your logic (searching for the "." in your header names), Record.Selectfields applies a filter to a row, using the output of list.select as trigger, and Record.Values gives you the values of the rows that where filtered.
- the "_" is needed when you use the each keyword, and represents the input value of this function. (in my case in the the List.Select function and the _ represents a item in the list)
Hope things are a little more clearer now 🙂
If this post helps or solves your problem, please mark it as solution (to help other users find useful content and to acknowledge the work of users that helped you)
Kudoes are nice too
Have fun
Jimmy