Forum Discussion
Pivot/ transpose
- 6 years ago
Anonymous
generic function:
// output let Source = Table.FromRecords(Json.Document(Binary.Decompress(Binary.FromText("i65WcvZ3cVWyUnIyMDBU0lEKTyxLVbIy1FFKSSwBMpS8SnMUDBWMDIwMgJLK+WmleZklxUpWprU62HUaoeo0IEGrMVyrY2m6ghGmTkMDVK1GONxrTNBSIxzuNbIkxVaEg4NTCxQMsbgYp7UmcK3+ySXY/Iqm0xiLV30TK4H6MLWa4dKK7Nc8BXMsXrWojQUA",BinaryEncoding.Base64),Compression.Deflate))), result = fnTrans(Source, {"CODE"}) in result // fnTrans (sTbl as table, groupList as list)=> let fn=(tbl as table)=> let fd = List.Skip(Table.ColumnNames(tbl)), nfd = groupList&List.TransformMany({1..Table.RowCount(tbl)}, each fd, (x,y)=>y&Text.From(x)) in Table.FromRows({Record.ToList((Table.SelectColumns(tbl, groupList)){0})&List.Combine(Table.ToRows(Table.SelectColumns(tbl, fd)))}, nfd), group = Table.Group(sTbl, groupList, {"t", fn}, 0)[t], result = Table.Combine(group) in result - 6 years ago
Here is another generic function
Value.ReplaceType((table as table) => let combineColumnName = Text.NewGuid() in Value.ReplaceType((primaryAttributeColumn as text, subAttributeColumns as list) as function => Value.ReplaceType((primaryAttributeValues as list) => let primaryValueColumnNameFn = each primaryAttributeColumn & "." & Text.From(_), primaryValueColumnNames = List.Transform(primaryAttributeValues, primaryValueColumnNameFn), attributeColumnToText = Table.TransformColumns(table, {primaryAttributeColumn, primaryValueColumnNameFn, type text}), mergeNewColumns = Table.CombineColumns(attributeColumnToText, subAttributeColumns, Value.ReplaceType(each Record.FromList(_, subAttributeColumns), type function(values as list) as Type.TableRow(Value.Type(Table.SelectColumns(#table(Value.Type(attributeColumnToText), {}), subAttributeColumns)))), combineColumnName), pivot = Table.Pivot(mergeNewColumns, primaryValueColumnNames, primaryAttributeColumn, combineColumnName), expandPivotColumn = List.Accumulate(primaryValueColumnNames, pivot, (t, c) => Table.ExpandRecordColumn(t, c, subAttributeColumns, List.Transform(subAttributeColumns, each c & "." & _))) in expandPivotColumn, type function (primaryAttributeValues as (type {text} meta [Documentation.AllowedValues = List.Distinct(Table.Column(table, primaryAttributeColumn))])) as {text} meta [Documentation.Description = "Pivot Multiple (Step 3)", Documentation.LongDescription = "Step 3: Confirm the primary attribute values"]), type function ( primaryAttributeColumn as (type text meta [Documentation.AllowedValues = Table.ColumnNames(table)]), subAttributeColumns as (type {text} meta [Documentation.AllowedValues = Table.ColumnNames(table)]) ) as function meta [Documentation.Description = "Pivot Multiple (Step 2)", Documentation.LongDescription = "Step 2: Choose the primary attribute, and the list of sub attributes"] ), (type function (table as table) as table) meta [Documentation.Description = "Pivot Multiple", Documentation.LongDescription = "Step 1: Choose a table to perform a multiple Pivot"] )
Thank you for the quick response!
I wanted to have this table as unique table (on the column CODE) as I am trying to merge/link these columns to another table that is also unique. I do not want the other table to have duplicate values for the code column.
Hope I explained it properly.
Hi Anonymous ,
It sounds like you need a dimension table. Create a table that contains only unique values of [Code], then relate this to your other tables that do not have unique values.
One way you can create a dimension table manually in Power Query is by using Enter Data, or you can do this is the data model by creating a new table with the following formula, but there are actually many ways to achieve this:
code = SUMMARIZE(aTable, aTable[code])
Then relate the tables as follows:
This will allow you to use both tables' data in calculations and visualisations using the dimension table as a common field between the two.
Pete