Forum Discussion
Add Multiple Columns from List that Have Standard Mathematical Operation
First, let me acknowledge that getting my result might be easier in DAX, but I would like to accomplish this in PQ for a host of reasons.
I have a DataTable below. The number of Fact columns that exist is variable, but I do have a List FactList[Fact Name] that contains each fact included in the DataTable
| Market | Product | Period | Fact 1 | Fact 2 | Fact 3 |
| Market 1 | Product 1 | Period 1 | |||
| Market 1 | Product 1 | Period 2 | |||
| Market 1 | Product 2 | Period 1 | |||
| Market 1 | Product 2 | Period 2 | |||
| Market 2 | Product 1 | Period 1 | |||
| Market 2 | Product 1 | Period 2 | |||
| Market 2 | Product 2 | Period 1 | |||
| Market 2 | Product 2 | Period 2 |
For each fact in FactList[Fact Name], I'd like to add 1 column to my DataTable. This column should subtract Period 1 - Period 2 for the given fact.
My attempt at figuring this out is as follows:
ListOfDimensions = Record.ToList([A = "Market", B = "Product", C = "Period"]),
ListOfFacts = FactList[FactName],
AddMergeColumn = Table.AddColumns(DataTable, "NonPeriodLookup", each Text.Combine({[Market],[Product]}, ""), type text),
ListOfNonFacts = List.Union({List.Select(ListOfDimensions , each _ <> "Period"), {"NonPerLookup"}})),
//Table with only Period 1 rows, facts names appended with " P1"
Period1 = Table.RemoveColumns(
Table.TransformColumnNames(
Table.SelectRows(AddMergeColumn , each [Period] = "Period 1"),
each if List.Contains(ListOfFacts, _) then _ & " P1" else _
),
"Period"
),
//Table with only Period 2 rows, facts names appended with " P2"
Period2 = Table.RemoveColumns(
Table.TransformColumnNames(
Table.SelectRows(AddMergeColumn , each [Period] = "Period 2"),
each if List.Contains(ListOfFacts, _) then _ & " P2" else _
),
"Period"
),
#"Merged Queries2" = Table.NestedJoin(Period1 ,{"NonPeriodLookup"},Period2, {"NonPeriodLookup"},"MergedP2",JoinKind.LeftOuter),
ListOfColumnsToExpand = Table.ColumnNames(Table.RemoveColumns(Period2, ListOfNonFacts)),
#"Expanded P2" = Table.ExpandTableColumn(#"Merged Queries2","MergedP2",ListOfColumnsToExpand),
AccumulateFacts =
List.Accumulate(ListOfFacts, #"Expanded P2",
(state, current) => Table.AddColumn(state, current & " Chg", each [P1] - [P2]
)
)
in
AccumulateFacts
This is my first real attempt at using List.Accumulate, but I think(?) I'm doing it correctly. The issue I'm running into is that I can't figure out how to refernce the two columns I want to subtract. In the above where I have [P1] - [P2], I really mean "[" & current & " P1]" - "[" & current & " P2]", but I don't know how to correctly reference columns dynmically.
Any help would be greatly appreciated. Thanks!
4 Replies
msmays5 Not sure if easier or harder, but definitely pretty simple in DAX. But, I know that ImkeF has solved this and probably edhans has as well. I'm sure they will be along shortly or maybe HotChilli , quite a few good Power Query people on the boards.
Just for reference (FYI only), one DAX method is the MTBF pattern. See my article on Mean Time Between Failure (MTBF) which uses EARLIER: http://community.powerbi.com/t5/Community-Blog/Mean-Time-Between-Failure-MTBF-and-Power-BI/ba-p/339586.
The basic pattern is:
Column =
VAR __Current = [Value]
VAR __PreviousDate = MAXX(FILTER('Table','Table'[Date] < EARLIER('Table'[Date])),[Date])
VAR __Previous = MAXX(FILTER('Table',[Date]=__PreviousDate),[Value])
RETURN
__Current - __PreviousHere's some code to make the provided algorithm work:
#"Added Index" = Table.AddIndexColumn(#"Expanded P2", "Index", 0, 1, Int64.Type), AccumulateFacts = List.Accumulate(ListOfFacts, #"Added Index", (state, current) => Table.AddColumn(state, current & " Chg", each Number.FromText(Table.Column(#"Added Index", current & " P1"){[Index]}) - Number.FromText(Table.Column(#"Added Index", current & " P2"){[Index]}) ) )I added an Index to get a quick way of identifying the row. Pretty sure the code can be made to work without it but I'll leave that as an exercise.
Also, the code provided by the OP didn't pass syntax check.
