Convert table with year rows and month columns to table with date column

[h4tt3n]

Hi,

I have got a table where rows represent years and columns represent months.

The table also contains a "Normal" row with averages per month, and a "Total" column with sums per year.

(The values are degree-days, and I need them for energy consumption calculations and optimization) 

I would like to give it a structure similar to a calendar- or time-series table, with a Year-Month / Date, and Value column, so I can relate it to my calendar table.

As I see it, we would need an extra column for "Normal", and an extra row per year for "Total", alternatively skip this value, since it can be easily calculated.

I have spent the last couple of hours testing different solutions, eg. crossjoining separate month and year based tables, without much progress. How would I do this?

Cheers, Mike

[Anonymous · Accepted solution]

If that's not what you want, spend another hour to publish a copyable table

let
    Origine = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("TZNLksMgDAXv4nUW1geQzpLK/a8xpgWDN6qUX/ml3Yjv99Jb7utzyc2s38o0pjMbszMHM5g5p/KuPu/+PhQKhY3CQWFSxfNmVJEO0khKhJKZqs5UbexCfbE5bEEVz3t/8ThVQYlSMlOFX7vvQhgEBoPKIek8H9AmDHyL6kzVYWvQDtKUXegvb8UG1eB56vIzq2DzvnieGTM1HJrELoQND+LtGAv+F7fLEqmSKqmRGqk96Srs50zL22KDRIqqLEEbTlWnaqaGQxu6CzFmUOFEAj83rrCqDSp82m1UDaqEKggzd2GcfaszXd6KDVfsnmZR4c1h41sch659F+axN+zsGNuobRxj7IOxn9ahzUHVTP3Z2CrU112IvfP/3mBb54hn475YFhWTdz32TdG6EXWy8EixwYlPkzpHvGHS2UlnGx2H7d43RetGvO5C7VudaXlbbJxvztRtpo7zBn+z56b8/gA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [year = _t, gen = _t, feb = _t, mar = _t, apr = _t, mag = _t, giu = _t, lug = _t, ago = _t, set = _t, ott = _t, nov = _t, dic = _t]),
    #"Modificato tipo" = Table.TransformColumnTypes(Origine,{{"year", Int64.Type}, {"gen", Int64.Type}, {"feb", Int64.Type}, {"mar", Int64.Type}, {"apr", Int64.Type}, {"mag", Int64.Type}, {"giu", Int64.Type}, {"lug", Int64.Type}, {"ago", Int64.Type}, {"set", Int64.Type}, {"ott", Int64.Type}, {"nov", Int64.Type}, {"dic", Int64.Type}}),
    #"Trasformate altre colonne tramite UnPivot" = Table.UnpivotOtherColumns(#"Modificato tipo", {"year"}, "Attributo", "Valore")
in
    #"Trasformate altre colonne tramite UnPivot"

[Vijay_A_Verma · Accepted solution]

See the working here - Open a blank query - Home - Advanced Editor - Remove everything from there and paste the below code to test (later on when you use the query on your dataset, you will have to change the source appropriately. If you have columns other than these, then delete Changed type step and do a Changed type for complete table from UI again)

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("XZA7DgIxDETvknqL2HFs5xJcYLUFPQiJ+xfkhdBQ7MifyRtrz7No1VqOItqmUpk5mn33w2OqJnUXdl5977IqilirUa5jAQVrByDxw4gnVpfd/399oahUp2mTmI00QoOdJ9FhUHPEOpx5rPMjx9TWUBuoN/IG9pDIhb293s/7g0gMIAjlIc/oZXERRsLhAu37p8TKdX0A", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Year = _t, January = _t, February = _t, March = _t, April = _t, May = _t, June = _t, July = _t, August = _t, September = _t, October = _t, November = _t, December = _t, Total = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"Year", type text}, {"January", Int64.Type}, {"February", Int64.Type}, {"March", Int64.Type}, {"April", Int64.Type}, {"May", Int64.Type}, {"June", Int64.Type}, {"July", Int64.Type}, {"August", Int64.Type}, {"September", Int64.Type}, {"October", Int64.Type}, {"November", Int64.Type}, {"December", Int64.Type}, {"Total", Int64.Type}}),
    #"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"Year"}, "Attribute", "Value"),
    #"Filtered Rows" = Table.SelectRows(#"Unpivoted Other Columns", each ([Year] <> "Normal")),
    #"Added Index" = Table.AddIndexColumn(#"Filtered Rows", "Index", 0, 1, Int64.Type),
    #"Added Custom" = Table.AddColumn(#"Added Index", "Date", each try Date.From("1"&[Attribute]&[Year]) otherwise [Year]&"-"&[Attribute]),
    Custom1 = Table.SelectRows(#"Unpivoted Other Columns", each ([Year] = "Normal")),
    #"Merged Queries" = Table.NestedJoin(#"Added Custom", {"Attribute"}, Custom1, {"Attribute"}, "Custom1", JoinKind.LeftOuter),
    #"Expanded Custom1" = Table.ExpandTableColumn(#"Merged Queries", "Custom1", {"Value"}, {"Value.1"}),
    #"Sorted Rows" = Table.Sort(#"Expanded Custom1",{{"Index", Order.Ascending}}),
    #"Removed Columns" = Table.RemoveColumns(#"Sorted Rows",{"Year", "Attribute", "Index"}),
    #"Reordered Columns" = Table.ReorderColumns(#"Removed Columns",{"Date", "Value", "Value.1"}),
    #"Renamed Columns" = Table.RenameColumns(#"Reordered Columns",{{"Value.1", "Normal"}})
in
    #"Renamed Columns"

[h4tt3n]

@Vijay_A_Verma 

Thank you for a very detailed answer. This has given me all the toos I need to solve the challenge.

Cheers, Mike

[Anonymous · Accepted solution]

If that's not what you want, spend another hour to publish a copyable table

let
    Origine = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("TZNLksMgDAXv4nUW1geQzpLK/a8xpgWDN6qUX/ml3Yjv99Jb7utzyc2s38o0pjMbszMHM5g5p/KuPu/+PhQKhY3CQWFSxfNmVJEO0khKhJKZqs5UbexCfbE5bEEVz3t/8ThVQYlSMlOFX7vvQhgEBoPKIek8H9AmDHyL6kzVYWvQDtKUXegvb8UG1eB56vIzq2DzvnieGTM1HJrELoQND+LtGAv+F7fLEqmSKqmRGqk96Srs50zL22KDRIqqLEEbTlWnaqaGQxu6CzFmUOFEAj83rrCqDSp82m1UDaqEKggzd2GcfaszXd6KDVfsnmZR4c1h41sch659F+axN+zsGNuobRxj7IOxn9ahzUHVTP3Z2CrU112IvfP/3mBb54hn475YFhWTdz32TdG6EXWy8EixwYlPkzpHvGHS2UlnGx2H7d43RetGvO5C7VudaXlbbJxvztRtpo7zBn+z56b8/gA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [year = _t, gen = _t, feb = _t, mar = _t, apr = _t, mag = _t, giu = _t, lug = _t, ago = _t, set = _t, ott = _t, nov = _t, dic = _t]),
    #"Modificato tipo" = Table.TransformColumnTypes(Origine,{{"year", Int64.Type}, {"gen", Int64.Type}, {"feb", Int64.Type}, {"mar", Int64.Type}, {"apr", Int64.Type}, {"mag", Int64.Type}, {"giu", Int64.Type}, {"lug", Int64.Type}, {"ago", Int64.Type}, {"set", Int64.Type}, {"ott", Int64.Type}, {"nov", Int64.Type}, {"dic", Int64.Type}}),
    #"Trasformate altre colonne tramite UnPivot" = Table.UnpivotOtherColumns(#"Modificato tipo", {"year"}, "Attributo", "Valore")
in
    #"Trasformate altre colonne tramite UnPivot"

[h4tt3n]

Hi @Anonymous 

That did the trick - Unpivot columns. Thank you!

In the future I will include a copyable table.

Cheers, Mike

[HotChilli]

I think you want to Unpivot the month columns (in Power Query).

You can then construct a Yr/Month column from the result.  This can be a date column if you want e.g. 1 feb 2011 (use Add column from examples if you are having trouble doing this) or a text column with your own design e.g. 2011/Feb

[h4tt3n]

Hi @HotChilli 

This is exactly what I have been attempting to do, but the details elude me a bit here. By unpivoting months, you also pivot years - the two are always at right angles to each other. Creating two separate month- and year columned tables and crossjoining them creates a mess.