Forum Discussion
Helper IIIOptimise Query for Calculating Number of Business Days
I have a table "DateTable" which has two columns:
- the first is every date between a start date (1st March 2019) and the current date
- the second is the number of business days that have occurred between that row and the first row (i.e. excluding weekends and holidays), where holidays are stored in a table "bank-holidays-dl".
A snippet of the table:
I set up the DateTable by creating a list of dates between the start and end date using List.Dates, converting it to a table, and then applying the following custom function to get the number of business days counting from the 01/03/2019:
(StartDate as date, EndDate as date) as number =>
let
ListDates = List.Dates(StartDate, Number.From(EndDate - StartDate), #duration(1, 0, 0, 0)),
RemoveWeekends = List.Select(ListDates, each Date.DayOfWeek(_, Day.Monday) < 5),
RemoveHolidays = List.RemoveItems(RemoveWeekends, #"bank-holidays-dl"),
CountDays = List.Count(RemoveHolidays)
in
CountDays
The function works as expected, however refreshing the data takes a long time. I was wondering if there could be any way to optimise this? Thank you!
The key to performance is to only calculate this column once instead of once for each row of your date table.
My approach would be to add an index column after filtering out weekends and holidays and then merge this back with the original dates.
let StartDate = #date(2019, 3, 1), EndDate = Date.From(DateTime.LocalNow()), ListDates = List.Dates(StartDate, Number.From(EndDate - StartDate), #duration(1, 0, 0, 0)), ListToTable = Table.FromList(ListDates, Splitter.SplitByNothing(), {"Date"}), DateTable = Table.TransformColumnTypes(ListToTable, {{"Date", type date}}), RemoveDays = Table.SelectRows(DateTable, each Date.DayOfWeek([Date], 1) < 5 and not List.Contains(#"bank-holidays-dl", AddIndex = Table.AddIndexColumn(RemoveDays, "Index", 1, 1, Int64.Type), MergeQueries = Table.NestedJoin(DateTable, {"Date"}, AddIndex, {"Date"}, "RemovedDays", JoinKind.LeftOuter), ExplandIndex = Table.ExpandTableColumn(MergeQueries, "RemovedDays", {"Index"}, {"WorkingDaysSinceStart"}), SortRows = Table.Sort(ExplandIndex,{{"Date", Order.Ascending}}), FillDown = Table.FillDown(SortRows, {"WorkingDaysSinceStart"}) in FillDown
3 Replies
- AlexisOlson
Super User
The key to performance is to only calculate this column once instead of once for each row of your date table.
My approach would be to add an index column after filtering out weekends and holidays and then merge this back with the original dates.
let StartDate = #date(2019, 3, 1), EndDate = Date.From(DateTime.LocalNow()), ListDates = List.Dates(StartDate, Number.From(EndDate - StartDate), #duration(1, 0, 0, 0)), ListToTable = Table.FromList(ListDates, Splitter.SplitByNothing(), {"Date"}), DateTable = Table.TransformColumnTypes(ListToTable, {{"Date", type date}}), RemoveDays = Table.SelectRows(DateTable, each Date.DayOfWeek([Date], 1) < 5 and not List.Contains(#"bank-holidays-dl", AddIndex = Table.AddIndexColumn(RemoveDays, "Index", 1, 1, Int64.Type), MergeQueries = Table.NestedJoin(DateTable, {"Date"}, AddIndex, {"Date"}, "RemovedDays", JoinKind.LeftOuter), ExplandIndex = Table.ExpandTableColumn(MergeQueries, "RemovedDays", {"Index"}, {"WorkingDaysSinceStart"}), SortRows = Table.Sort(ExplandIndex,{{"Date", Order.Ascending}}), FillDown = Table.FillDown(SortRows, {"WorkingDaysSinceStart"}) in FillDown
oliverblaneThis worked perfectly, thank you!
CNENFRNLCommunity Champion
(StartDate as date, EndDate as date) => List.Accumulate( List.Dates(StartDate, Duration.TotalDays(EndDate - StartDate), #duration(1, 0, 0, 0)), 0, (s,c) => if Date.DayOfWeek(c, Day.Monday) >= 5 or List.Contains(#"bank-holidays-dl", c) then s else s+1 )