Performance of counting overlapped datetime

Hi hirokichi 

Groupby is slow, so did those 2 queries above achieve what you wanted? How about buffer the table in Query 2?

let
    source = Csv.Document(File.Contents("C:\SampleTime.csv"),[Delimiter=",", Columns=3, Encoding=932, QuoteStyle=QuoteStyle.None]),
    header = Table.PromoteHeaders(source, [PromoteAllScalars=true]),
    convertType =Table.Buffer( Table.TransformColumnTypes(header,{{"Id", Int64.Type}, {"StartDate", type datetime}, {"EndDate", type datetime}})),
    addColumn = Table.AddColumn(convertType, "Overlap", each Table.RowCount(Table.SelectRows(convertType, (x) => x[StartDate] <= [StartDate] and [StartDate] <= x[EndDate])))
in
    addColumn

try query 3

let
    source = Csv.Document(File.Contents("C:\SampleTime.csv"),[Delimiter=",", Columns=3, Encoding=932, QuoteStyle=QuoteStyle.None]),
    header = Table.PromoteHeaders(source, [PromoteAllScalars=true]),
    convertType = Table.Buffer( Table.TransformColumnTypes(header,{{"Id", Int64.Type}, {"StartDate", type datetime}, {"EndDate", type datetime}})),
    Custom = List.Buffer( Table.AddColumn(convertType, "Custom", each {[StartDate], [EndDate]})[Custom]),
    #"Added Custom1" = Table.AddColumn(convertType, "Overlap", (x)=> List.Count(List.Select( List.Transform(Custom, each _{0} <=x[StartDate] and _{1}>=x[StartDate]), each _=true)))
in
    #"Added Custom1"