Forum Discussion
Replacing a column with calculation, performing a column lookup dynamically
Add a custom step with the following. Replace #"Changed Type" with the name of your last step (that is the default value when you add a custom step).
= Table.FromRecords(Table.TransformRows(#"Changed Type", each _ & Record.FromList(List.Transform(List.Select(Record.FieldNames(_), each Text.StartsWith(_, "Score_")), (name) => Record.Field(_, name) * Record.Field(_, "Weight_" & Text.AfterDelimiter(name, "Score_"))),List.Select(Record.FieldNames(_), each Text.StartsWith(_, "Score_")))), Value.Type(#"Changed Type"))
Try this Anonymous
It turns this table:
into this:
What it does is:
- Normalizes the data by unpivoting all of the score/weight columns
- Splits the score_1 into two fields (score, 1)
- Pivots the category that has the score/weight lables so they are now in columns
- adds a custom column that does the math. If weight is 0 it returns 0 - otherwise you'd get an infinity answer which you likely do not want.
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("TY/NCsIwEITfJecQmt/Wo4gXwScIPSwYsNgfaXx/TKar5LIMs19mNjGKGy0pCymMMmV2SpeplYc20BbailFGcaf9BdP8AaN6aAd/YL/Cl217p70YDssjO/DT0HQe2dd5nhJCTuxK0XO7Ax1YI/xJOSHvILp6smXENOEe+Hmh9UHFGPhv9RbPuAVu+S7gK+UP5YmaBo0C1xTYX8H4BQ==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Name = _t, Score_1 = _t, Weight_1 = _t, Score_2 = _t, Weight_2 = _t, Score_3 = _t, Weight_3 = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Score_1", Int64.Type}, {"Weight_1", Int64.Type}, {"Score_2", Int64.Type}, {"Weight_2", Int64.Type}, {"Score_3", Int64.Type}, {"Weight_3", Int64.Type}}),
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"Name"}, "Category", "Value"),
#"Split Column by Delimiter" = Table.SplitColumn(#"Unpivoted Other Columns", "Category", Splitter.SplitTextByDelimiter("_", QuoteStyle.Csv), {"Category", "Version"}),
#"Pivoted Column" = Table.Pivot(#"Split Column by Delimiter", List.Distinct(#"Split Column by Delimiter"[Category]), "Category", "Value"),
#"Added Custom" = Table.AddColumn(#"Pivoted Column", "New Value", each if [Weight] = 0 then 0
else 100 - [Score] / [Weight], Int64.Type
)
in
#"Added Custom"From there you can transform as desired to return only what you need to the DAX model for visuals.
How to use M code provided in a blank query:
1) In Power Query, select New Source, then Blank Query
2) On the Home ribbon, select "Advanced Editor" button
3) Remove everything you see, then paste the M code I've given you in that box.
4) Press Done
5) See this article if you need help using this M code in your model.
Thanks Edhans, so in this section, if I have 299 columns, I would need to put in all of them? In the end I will need to unpivot anyway as that is how I need to report it so that works out.
BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Name = _t, Score_1 = _t, Weight_1 = _t, Score_2 = _t, Weight_2 = _t, Score_3 = _t, Weight_3 = _t]),
Also, can I just replace the score column instead of creating a new one? I have to actually take this and do a merge query to do a compare so the column name needs to stay the same.
No. that first line is what I pasted in using the Enter Data feature.
Your source would the data connection to your 299 column file. My code should work on all columns - that is be beauty of UnPivot. See the link I posted above on how to integrate my code sample into your actual table. It has images on how to merge the code, and a link at the bottom of that post to a video Imke made for this as well.
You can then remove unneeded columns and rename columns to suit your needs for subsequent merges.Please post back if you have any specific questions on being unable to do the code merge and I'll assist. I'd need your M code to do it though. Otherwise, if you can mark my post as the solution it would be appreciated, and also let others know this is resolved and there is an answer for future users searching similar scenarios.
Thanks edhans , artemus . I was trying to add that last step, having issues but this looks like the solution that I am looking for.
The last step gave me a "comma token expected" error.
let Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("TY/NCsIwEITfJecQmt/Wo4gXwScIPSwYsNgfaXx/TKar5LIMs19mNjGKGy0pCymMMmV2SpeplYc20BbailFGcaf9BdP8AaN6aAd/YL/Cl217p70YDssjO/DT0HQe2dd5nhJCTuxK0XO7Ax1YI/xJOSHvILp6smXENOEe+Hmh9UHFGPhv9RbPuAVu+S7gK+UP5YmaBo0C1xTYX8H4BQ==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Name = _t, Score_1 = _t, Weight_1 = _t, Score_2 = _t, Weight_2 = _t, Score_3 = _t, Weight_3 = _t]), #"Changed Type" = Table.TransformColumnTypes(Source,{{"Score_1", type number}, {"Weight_1", type number}, {"Score_2", Int64.Type}, {"Weight_2", Int64.Type}, {"Score_3", Int64.Type}, {"Weight_3", Int64.Type}}), #"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"Name"}, "Category", "Value"), #"Split Column by Delimiter" = Table.SplitColumn(#"Unpivoted Other Columns", "Category", Splitter.SplitTextByDelimiter("_", QuoteStyle.Csv), {"Category", "Version"}), #"Pivoted Column" = Table.Pivot(#"Split Column by Delimiter", List.Distinct(#"Split Column by Delimiter"[Category]), "Category", "Value"), #"Added Custom" = Table.AddColumn(#"Pivoted Column", "New Value", each if [Weight] = 0 then 0 else 100 - [Score] / [Weight], Int64.Type ), #"Added Custom1" = Table.FromRecords(Table.TransformRows(#"Added Custom", each _ & Record.FromList(List.Transform(List.Select(Record.FieldNames(_), each Text.StartsWith(_, "Score_")), (name) => Record.Field(_, name) * Record.Field(_, "Weight_" & Text.AfterDelimiter(name, "Score_"))),List.Select(Record.FieldNames(_), each Text.StartsWith(_, "Score_")))), Value.Type(#"Added Custom"))) in #"Added Custom1"