Excel MVPs angriper datarensingsproblemet i Power Query - Excel Tips

Innholdsfortegnelse

Merk

Dette er en av en serie artikler som beskriver løsninger som er sendt inn for Podcast 2316-utfordringen.

Excel MVP Oz Du Soleil fra Excel on Fire-kanalen på YouTube nevnte Brazilian Bull Rider Kaique Pachecho. Oz var den første personen som la merke til at jeg gikk langsomt for å legge til de fire kvartalene.

Ozs video er:
https://www.youtube.com/watch?v=OluZlF44PNI

Hans kode er:

let Source = Excel.CurrentWorkbook()((Name="UglyData"))(Content), #"Removed Columns" = Table.RemoveColumns(Source,("Column2", "Column3", "Column4", "Column5", "Column6")), #"Transposed Table" = Table.Transpose(#"Removed Columns"), #"Promoted Headers" = Table.PromoteHeaders(#"Transposed Table", (PromoteAllScalars=true)), #"Changed Type" = Table.TransformColumnTypes(#"Promoted Headers",(("Category Description", type text), ("Administrative", type number), ("Holiday", Int64.Type), ("PTO/LOA/Jury Duty", Int64.Type), ("Project A", type number), ("Project B", type number), ("Project C", type number))), #"Added Conditional Column" = Table.AddColumn(#"Changed Type", "Custom", each if (Category Description) = "Q1" then null else if (Category Description) = "Q2" then null else if (Category Description) = "Q3" then null else if (Category Description) = "Q4" then null else (Category Description)), #"Filled Down" = Table.FillDown(#"Added Conditional Column",("Custom")), #"Renamed Columns" = Table.RenameColumns(#"Filled Down",(("Custom", "Names"))), #"Filtered Rows" = Table.SelectRows(#"Renamed Columns", each (Category Description) = "Q1" or (Category Description) = "Q2" or (Category Description) = "Q3" or (Category Description) = "Q4"), #"Reordered Columns" = Table.ReorderColumns(#"Filtered Rows",("Names", "Category Description", "Administrative", "Holiday", "PTO/LOA/Jury Duty", "Project A", "Project B", "Project C")), #"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Reordered Columns", ("Names", "Category Description"), "Attribute", "Value"), #"Pivoted Column" = Table.Pivot(#"Unpivoted Other Columns", List.Distinct(#"Unpivoted Other Columns"(#"Category Description")), "Category Description", "Value", List.Sum), #"Inserted Sum" = Table.AddColumn(#"Pivoted Column", "Addition", each List.Sum(((Q1), (Q2), (Q3), (Q4))), type number), #"Renamed Columns1" = Table.RenameColumns(#"Inserted Sum",(("Addition", "TOTAL"))) in #"Renamed Columns1"

En annen løsning, denne fra Excel MVP John MacDougall.

  • John var den første til å si at ved å slette de to ekstra trinnene Power Query la til, eliminerer du de merkelige suffiksen på duplikatene Q1 Q2 Q3 Q4.
  • John brukte en indekskolonne tidlig som skulle brukes til slutt for sortering. Men - John sammenkoblet indekskolonnen etter kategoribeskrivelsen. Han brukte et loddrett rørkarakter | slik at han kunne bryte ut dataene senere.
  • John skrev sin betingede kolonne som en egendefinert kolonne i stedet for å bruke grensesnittet for betinget kolonne.
Betinget kolonne som en tilpasset kolonne

Se Johns video her:
https://www.youtube.com/watch?v=Dqmb6SEJDXI

Excel MVP Ken Puls, medforfatter av M er for (Data) Monkey book sendt i tre løsninger. Hans betingede kolonne er trolig den korteste.

Men Ken's foretrukne løsning ignorerer det opprinnelige spørsmålet. I stedet for å lage tabellen i Power Query, oppretter han et svingbart datasett i Power Query og avsluttes deretter med en pivottabell.

Kens siste forhåndsvisning i Power Query ser slik ut:

Svingbart datasett

Her er Kens kode:

let Source = Excel.CurrentWorkbook()((Name="UglyData"))(Content), #"Promoted Headers" = Table.PromoteHeaders(Source, (PromoteAllScalars=true)), #"Changed Type" = Table.TransformColumnTypes(#"Promoted Headers",(("Category Description", type text), ("Dept. Total", type number), ("Q1", type number), ("Q2", type number), ("Q3", type number), ("Q4", Int64.Type), ("Employee 1", type number), ("Q1_1", type number), ("Q2_2", type number), ("Q3_3", Int64.Type), ("Q4_4", Int64.Type), ("Employee 2", Int64.Type), ("Q1_5", Int64.Type), ("Q2_6", Int64.Type), ("Q3_7", Int64.Type), ("Q4_8", Int64.Type), ("Employee 3", Int64.Type), ("Q1_9", Int64.Type), ("Q2_10", Int64.Type), ("Q3_11", Int64.Type), ("Q4_12", Int64.Type), ("Employee 4", type number), ("Q1_13", type number), ("Q2_14", type number), ("Q3_15", type number), ("Q4_16", Int64.Type))), #"Removed Columns" = Table.RemoveColumns(#"Changed Type",("Dept. Total", "Q1", "Q2", "Q3", "Q4")), #"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Removed Columns", ("Category Description"), "Attribute", "Value"), #"Added Conditional Column" = Table.AddColumn(#"Unpivoted Other Columns", "Employee", each if Text.Contains((Attribute), "_") then null else (Attribute)), #"Filled Down" = Table.FillDown(#"Added Conditional Column",("Employee")), #"Split Column by Delimiter" = Table.SplitColumn(#"Filled Down", "Attribute", Splitter.SplitTextByEachDelimiter(("_"), QuoteStyle.Csv, false), ("Attribute.1", "Attribute.2")), #"Changed Type1" = Table.TransformColumnTypes(#"Split Column by Delimiter",(("Attribute.1", type text), ("Attribute.2", Int64.Type))), #"Filtered Rows" = Table.SelectRows(#"Changed Type1", each ((Attribute.2) null)), #"Removed Columns1" = Table.RemoveColumns(#"Filtered Rows",("Attribute.2")), #"Renamed Columns" = Table.RenameColumns(#"Removed Columns1",(("Attribute.1", "Quarter"), ("Value", "Amount"))), #"Changed Type2" = Table.TransformColumnTypes(#"Renamed Columns",(("Category Description", type text), ("Quarter", type text), ("Amount", type number), ("Employee", type text))) in #"Changed Type2"

Etter å ha opprettet dette spørsmålet bare som en tilkobling, bruker han deretter en pivottabell for å opprette den endelige rapporten.

Sluttrapport med pivottabell

Løsninger fra andre MVPer:

  • Wyn Hopkins-koden er her: Power Query: Dealing with Multiple Identical Headers.
  • Mike Girvins kode er her: Power Query: Extracting Left 2 Characters From a Column.
  • Roger Goviers formelløsning er her: Formula Solutions.

Gå tilbake til hovedsiden for Podcast 2316-utfordringen.

Les neste artikkel i denne serien: Power Query: Beyond the User Interface: Table.Split and More.

Interessante artikler...