Forum Discussion

MalAttari's avatar
MalAttari
Frequent Visitor
1 year ago
Solved

Changing consecutive dates to from dates

Hello  I have following table  Person Date Assignment Jack 01-Jan Support Jack 02-Jan Support Jack 03-Jan Support Jack 04-Jan Analysis Jack 05-Jan Analysis Jack...
  • DataNinja777's avatar
    1 year ago

    Hi MalAttari ,

     

    To achieve the desired transformation in Power Query, start by loading your table into Power Query and sorting it by the Person and Date columns in ascending order. After sorting, you need to identify consecutive rows with the same Assignment while ensuring that dates are sequential.

    First, create an Index column to assist in identifying groups of consecutive assignments. To do this, go to the Add Column tab and select Index Column (From 0). Next, add a Custom Column to calculate a GroupID that identifies each consecutive block of the same assignment. You can do this using the following steps.

    Go to Add Column > Custom Column and use the formula below:

     

     

    let
        // Load your table
        Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
    
        // Sort the table by Person and Date
        SortedTable = Table.Sort(Source, {{"Person", Order.Ascending}, {"Date", Order.Ascending}}),
    
        // Buffer the table to improve performance and prevent referencing issues
        BufferedTable = Table.Buffer(SortedTable),
    
        // Add an Index column to help track row positions
        IndexedTable = Table.AddIndexColumn(BufferedTable, "Index", 0, 1, Int64.Type),
    
        // Add a GroupID column to identify consecutive assignments
        GroupedTable = Table.AddColumn(
            IndexedTable,
            "GroupID",
            each if [Index] = 0 then 0
            else if [Assignment] = BufferedTable{[Index]-1}[Assignment] 
                 and [Person] = BufferedTable{[Index]-1}[Person] 
                 and Date.AddDays(BufferedTable{[Index]-1}[Date], 1) = [Date] 
            then BufferedTable{[Index]-1}[GroupID]
            else BufferedTable{[Index]-1}[GroupID] + 1
        ),
    
        // Group by Person and GroupID, and calculate From and To dates
        GroupedResult = Table.Group(
            GroupedTable,
            {"Person", "GroupID"},
            {
                {"From", each List.Min([Date]), type date},
                {"To", each List.Max([Date]), type date},
                {"Assignment", each List.First([Assignment]), type text}
            }
        ),
    
        // Remove the GroupID column
        FinalTable = Table.RemoveColumns(GroupedResult, {"GroupID"})
    in
        FinalTable
    

     

     

     

    Best regards,

  • slorin's avatar
    1 year ago

    Hi MalAttari 

     

    let
    Source = Your_Source,
    Index = Table.AddIndexColumn(Your_Source, "Index", 0, 1, Int64.Type),
    Group = Table.Group(
    Index,
    {"Person", "Assignment", "Date", "Index"},
    {{"From", each List.Min([Date]), type nullable date}, {"To", each List.Max([Date]), type nullable date}},
    GroupKind.Local,
    (x,y) => Number.From((y[Index]-x[Index]) <> Number.From(Duration.Days(y[Date]-x[Date]))
    or y[Assignment] <> x[Assignment])),
    Columns = Table.SelectColumns(Group,{"Person", "From", "To", "Assignment"})
    in
    Columns

    Stéphane

  • slorin's avatar
    slorin
    1 year ago

    speedramps 

    You have to adapt this code

    let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45W8kpMzlbSUTIw1AciIwMjUyAnuLSgIL+oRClWByFvREDemIC8CZK8Y15iTmVxZjGKAlNCCiwIKDDE6sRYAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Person = _t, Date = _t, Assignment = _t]),
    TypeDate = Table.TransformColumnTypes(Source, {{"Date", type date}}, "fr-FR"),
    Index = Table.AddIndexColumn(TypeDate, "Index", 0, 1, Int64.Type),
    Group = Table.Group(Index, {"Person", "Assignment", "Date", "Index"}, {{"From", each List.Min([Date]), type nullable date}, {"To", each List.Max([Date]), type nullable date}}, GroupKind.Local, (x,y) => Number.From((y[Index]-x[Index])<>Number.From(Duration.Days(y[Date]-x[Date])) or y[Assignment]<>x[Assignment])),
    Columns = Table.SelectColumns(Group,{"Person", "From", "To", "Assignment"})
    in
    Columns

     Stéphane

  • dufoq3's avatar
    1 year ago

    Hi MalAttari, another solution:

     

    Output

    let
        Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45W8kpMzlbSUTIw1PVKzAMygksLCvKLSpRidRByRnjkjPHImUDlHPMScyqLM4tRJE3xSVrgkTTEcE4sAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Person = _t, Date = _t, Assignment = _t]),
        AddedIndex = Table.AddIndexColumn(Source, "Index", 0, 1, Int64.Type),
        GroupedRows = Table.Group(AddedIndex, {"Person", "Date", "Assignment", "Index"}, {{"T", each let a = Table.First(_), b = [Person = a[Person], From = a[Date], To = Table.Last(_)[Date], Assignment = a[Assignment]] in b, type table}}, 0,
            (x,y)=> Byte.From( y[Person] <> x[Person] or y[Assignment] <> x[Assignment] or Duration.TotalDays(Date.From(y[Date]) - Date.From(x[Date]))  > (y[Index] - x[Index]) ) ),
        CombinedT = Table.FromRecords(GroupedRows[T])
    in
        CombinedT