Forum Discussion
ADF pivot vs fabric dataflow pivot
I created the helper row by using Enter Data. Then I copied the generated M code into the main query manually.
But you can do it easier:
You can create it as a separate query by using Enter Data, and then appending that query to your main query by using the menu option for Append queries.
Something like this?
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMjQyNlHSUUrKz882BNK5+WWpual5JSB2cUlRYklqeiWInZJaXJBYkpyhkAIUAvINjfQMLPWMDIyMwRx9A0t9KAeCDA0M9AyMQDoNlGJ1aGYNCNHEeJD7DU0N4bbQ1BejgUWsLxAhNUitKcksyUmF2YGMyIxjnOaRGJlYzKEo1ka2P6mdPsg3L7GoKLMsMQfJRCMLoFF61ChfiDKbxPghYCZFcTUaFqS4l5ZpDt3sWAA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [venture_number = _t, book = _t, movement_name = _t, strategy = _t, name = _t, actual_date = _t, estimated_date = _t, gross = _t, net = _t, actual = _t, percentage = _t]),
#"Changed column type" = Table.TransformColumnTypes(Source, {{"venture_number", Int64.Type}, {"book", type text}, {"movement_name", type text}, {"strategy", type text}, {"name", type text}, {"gross", type number}, {"net", type number}, {"actual", type number}, {"percentage", Int64.Type}}),
#"Changed column type with locale 1" = Table.TransformColumnTypes(#"Changed column type", {{"actual_date", type date}, {"estimated_date", type date}}, "nb-NO"),
#"Removed duplicates" = Table.Distinct(#"Changed column type with locale 1", {"venture_number", "book", "movement_name", "strategy", "name", "actual_date", "estimated_date", "gross", "net", "actual", "percentage"}),
#"Grouped rows" = Table.Group(#"Removed duplicates", {"venture_number", "book", "movement_name", "strategy", "name"}, {{"actual_date", each List.Max([actual_date]), type nullable date}, {"estimated_date", each List.Max([estimated_date]), type nullable text}, {"gross", each List.Max([gross]), type nullable number}, {"net", each List.Max([net]), type nullable number}, {"actual", each List.Max([net]), type nullable number}, {"percentage", each List.Max([percentage]), type nullable Int64.Type}}),
#"Unpivoted only selected columns" = Table.Unpivot(#"Grouped rows", {"actual_date", "estimated_date"}, "Attribute", "Value"),
#"Changed column type with locale" = Table.TransformColumnTypes(#"Unpivoted only selected columns", {{"Value", type date}}, "nb-NO"),
#"Merged columns" = Table.CombineColumns(#"Changed column type with locale", {"name", "Attribute"}, Combiner.CombineTextByDelimiter(" ", QuoteStyle.None), "Merged"),
#"Pivoted column" = Table.Pivot(Table.TransformColumnTypes(#"Merged columns", {{"Merged", type text}}), List.Distinct(Table.TransformColumnTypes(#"Merged columns", {{"Merged", type text}})[Merged]), "Merged", "Value")
in
#"Pivoted column"
How do I use your code?
Can you guide me to create this ib the dstafkow
Tgankyiu
Inside the Dataflow Gen2, you can click Get Data -> Blank query, and then paste the code inside there, then click Next.
I am now implementing your suggestion on the whole actual data. but not sure why the remove duplicates only returns one row whereas it should return four rows similar to the example I sent you
In the remove duplicates step, you need to specify which columns shall be considered when comparing the rows.
If the difference between the rows exists in some column which is not specified in the removed duplicates step, then it will not detect that there is a difference between those rows.
So you will need to specify which columns to check.
Currently, it is only checking the columns which are mentioned here:
{"venture_number", "book", "movement_name", "strategy", "name", "actual_date", "estimated_date", "gross", "net", "actual", "percentage"}