Forum Discussion
Selected Value dax Query
A measure doesn't create rows so you won't be able to have those tabular lists. If the calculated table on your end, doesnt give you the correct order, you can just create two extra columns to kind of select the should be column values.
Sort2 = IFERROR ( VALUE ( NewTable[Sort] ), VALUE ( NewTable[Carrier] ) ) Carrier2 = IF ( NewTable[Sort] = NewTable[Type], NewTable[Type], NewTable[Carrier] )
The sorting should be the same as the corresponding Carrier.
The shorting should be like this.
You're better off doing this in M that in DAX. Below is the updated DAX calculated table formula but this returns the wrong order and won't change even if created as a new table.
NewTable_ =
VAR __TYPE =
DISTINCT ( OriginalTable[Type] )
VAR __TYPE2 =
DISTINCT (
SELECTCOLUMNS (
OriginalTable,
"Type", [Type],
"Carrier", [Type],
"Sort", [Index]
)
)
VAR __CROSSJOINED =
SELECTCOLUMNS (
FILTER (
CROSSJOIN (
__TYPE,
SELECTCOLUMNS (
OriginalTable,
"Type2", [Type],
"Carrier", [Carrier],
"Sort", [Index]
)
),
[Type] <> [Type2]
),
"Type", [Type],
"Sort", [Sort],
"Carrier", [Carrier]
)
VAR __LETTER_CARRIER =
ADDCOLUMNS ( __TYPE, "Carrier", [Type] )
RETURN
UNION ( __CROSSJOINED, __TYPE2 )
Here's the M code
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WclTSUXJOLCrKTC1SMFSK1YlWckISMQKLOCOJGINFXJBETMAirkgipmARNyQRM6XYWAA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Type = _t, Carrier = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Type", type text}, {"Carrier", type text}}),
#"Added Index" = Table.AddIndexColumn(#"Changed Type", "Index", 1, 1, Int64.Type),
#"Table from Previous Step" = Table.AddColumn(#"Added Index", "Previous Step", each #"Added Index"),
#"Expanded Previous Step" = Table.ExpandTableColumn(#"Table from Previous Step", "Previous Step", {"Type", "Carrier", "Index"}, {"Previous Step.Type", "Previous Step.Carrier", "Previous Step.Index"}),
#"Filtered Rows" = Table.SelectRows(#"Expanded Previous Step", each [Previous Step.Type] <> [Type]),
#"Removed Other Columns" = Table.SelectColumns(#"Filtered Rows",{"Type", "Previous Step.Carrier", "Previous Step.Index"}),
#"Renamed Columns" = Table.RenameColumns(#"Removed Other Columns",{{"Previous Step.Index", "Index"}}),
#"Appended Added Index to the Prev Step" = Table.Combine({#"Added Index", #"Renamed Columns"}),
#"Removed Columns" = Table.RemoveColumns(#"Appended Added Index to the Prev Step",{"Carrier"}),
#"Added Custom" = Table.AddColumn(#"Removed Columns", "Carrier", each if [Previous Step.Carrier] = null then [Type] else [Previous Step.Carrier]),
#"Removed Columns1" = Table.RemoveColumns(#"Added Custom",{"Previous Step.Carrier"}),
#"Reordered Columns" = Table.ReorderColumns(#"Removed Columns1",{"Type", "Carrier", "Index"}),
#"Changed Type1" = Table.TransformColumnTypes(#"Reordered Columns",{{"Carrier", type text}}),
#"Renamed Columns1" = Table.RenameColumns(#"Changed Type1",{{"Index", "Sort"}})
in
#"Renamed Columns1"
I think i am havign trouble using this concept to incorporate to my existing data model. Is it possible to use it as a measure instead?
A measure doesn't create rows so you won't be able to have those tabular lists. If the calculated table on your end, doesnt give you the correct order, you can just create two extra columns to kind of select the should be column values.
Sort2 = IFERROR ( VALUE ( NewTable[Sort] ), VALUE ( NewTable[Carrier] ) ) Carrier2 = IF ( NewTable[Sort] = NewTable[Type], NewTable[Type], NewTable[Carrier] )The problem is i am using this table as dim to be connected to a fact table. so the relationship is many to many.. and also only A has data when it is connected becasue in the fact table the type is only shown. not the corresponding column which is the carrier. For example :