Forum Discussion

Mike1983's avatar
Mike1983
Regular Visitor
2 years ago
Solved

How to count consecutive date ranges in Power Query M?

Hello! Hopefully someone can help me!

 

I have a table of absence start and end dates (300k rows). Where absences are consecutive (previous end date is the day before the next start date), I want to capture the first of the consecutive dates in that series as Consecutive Abs Start. It could be any number of absences in a consecutive series. Please see the desired result below. I want to do this in PowerQuery M (instead of PowerBI DAX front end) so I can use the result further in my transformations.

 

If anyone can help me that would be great! Thanks.  Mike

 

 

PersonAbs StartAbs EndConsecutive Abs Start (desired result)
2000000101.01.202402.01.202401.01.2024
2000000115.01.202417.01.202415.01.2024
2000000118.01.202419.01.202415.01.2024
2000000120.01.202420.01.202415.01.2024
2000000121.01.202421.01.202415.01.2024
2000000122.01.202422.01.202415.01.2024
2000000123.01.202423.01.202415.01.2024
2000000101.02.202401.02.202401.02.2024
2000000115.02.202415.02.202415.02.2024
2000000116.02.202416.02.202415.02.2024
2000000101.03.202401.03.202401.03.2024

 

 

 

Need Help
  • dufoq3's avatar
    dufoq3
    2 years ago

    Hi Mike1983,

     

    Result:

     

    let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("ddDRCoAgDIXhd/FaYjuW1bOI7/8atgg7jSbe/PDBdK0lyH005SS6XBeC1QIzev4w3YjpHrKD2RkxCDEKz/htFJ6BWfgFFGYlYrYQzIW88bOQySg8q8xqxGxO4aFP9D4A", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Person = _t, #"Abs Start" = _t, #"Abs End" = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"Abs Start", type date}, {"Abs End", type date}}, "sk-SK"),
    #"Added Index1" = Table.AddIndexColumn(#"Changed Type", "IndexHelper", 0, 1, Int64.Type),
    #"Added Dates" = Table.AddColumn(#"Added Index1", "Dates", each List.Dates([Abs Start], Duration.TotalDays([Abs End] - [Abs Start])+1, #duration(1,0,0,0)), type list),
    #"Expanded Dates" = Table.ExpandListColumn(#"Added Dates", "Dates"),
    #"Added Index" = Table.AddIndexColumn(#"Expanded Dates", "Index", 0, 1, Int64.Type),
    #"Grouped Rows" = Table.Group(#"Added Index",
        {"Person", "Dates", "Index"}, { {"All", each _, type table}, {" Consecutive Abs Start", each List.Min([Dates]), type date}},
        GroupKind.Local,
        (s,c)=> Byte.From(c[Index]-s[Index] <> Duration.Days(c[Dates]-s[Dates]))),
    #"Expanded All" = Table.ExpandTableColumn(#"Grouped Rows", "All", {"Abs Start", "Abs End", "IndexHelper"}, {"Abs Start", "Abs End", "IndexHelper"}),
    #"Removed Duplicates" = Table.Distinct(#"Expanded All", {"IndexHelper"}),
    #"Removed Columns" = Table.RemoveColumns(#"Removed Duplicates",{"IndexHelper", "Dates", "Index"})
in
    #"Removed Columns"

     

3 Replies

  • dufoq3's avatar
    dufoq3
    Community Champion
    2 years ago

    Hi Mike1983,

     

    Result:

     

    let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("ddDRCoAgDIXhd/FaYjuW1bOI7/8atgg7jSbe/PDBdK0lyH005SS6XBeC1QIzev4w3YjpHrKD2RkxCDEKz/htFJ6BWfgFFGYlYrYQzIW88bOQySg8q8xqxGxO4aFP9D4A", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Person = _t, #"Abs Start" = _t, #"Abs End" = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"Abs Start", type date}, {"Abs End", type date}}, "sk-SK"),
    #"Added Index1" = Table.AddIndexColumn(#"Changed Type", "IndexHelper", 0, 1, Int64.Type),
    #"Added Dates" = Table.AddColumn(#"Added Index1", "Dates", each List.Dates([Abs Start], Duration.TotalDays([Abs End] - [Abs Start])+1, #duration(1,0,0,0)), type list),
    #"Expanded Dates" = Table.ExpandListColumn(#"Added Dates", "Dates"),
    #"Added Index" = Table.AddIndexColumn(#"Expanded Dates", "Index", 0, 1, Int64.Type),
    #"Grouped Rows" = Table.Group(#"Added Index",
        {"Person", "Dates", "Index"}, { {"All", each _, type table}, {" Consecutive Abs Start", each List.Min([Dates]), type date}},
        GroupKind.Local,
        (s,c)=> Byte.From(c[Index]-s[Index] <> Duration.Days(c[Dates]-s[Dates]))),
    #"Expanded All" = Table.ExpandTableColumn(#"Grouped Rows", "All", {"Abs Start", "Abs End", "IndexHelper"}, {"Abs Start", "Abs End", "IndexHelper"}),
    #"Removed Duplicates" = Table.Distinct(#"Expanded All", {"IndexHelper"}),
    #"Removed Columns" = Table.RemoveColumns(#"Removed Duplicates",{"IndexHelper", "Dates", "Index"})
in
    #"Removed Columns"

     

    • Mike1983's avatar
      Mike1983
      Regular Visitor
      2 years ago

      Fantastic, thank you. I got there in the end but I used about 10 times as many steps, this is certainly much more elegant. 

      • dufoq3's avatar
        dufoq3
        Community Champion
        2 years ago

        You're welcome Mike. Enjoy. Don't forget that your data needs to be sorted by Abs Start.