How to handle mixed data format in the same column

[1tntla]

Hello guys,

There are 10 files in a folder.

They are all CSVs

They all have the same number of headers, in the same order, but different header names (eg: one is in Franch and the other is in Eng)

Issue is that when I combine all of them into a single table in PowerBI, the "Date" column has mixed formatting based on the region.

Example:

yyyy/mm/dd

yy/mm/dd

mm/dd/yyyy

etc. 

I could import them separately and correctly format the dates one by one and then combine them. However, is there a way to do this with all of them combined first?

Cheers,

Rok

[MarcelBeug]

Somehow you need to indicate the date format.

The code below contains a function embedded in sample code from an Excel file where datestrings are converted to dates based on an associated 3-letter sequence string (dmy, mdy, ymd or ydm, case insensitive) and date delimiters one of: slash (/), dash (-), period (.).

The year part must have 2 or 4 digits; in case of 2 digits it is assumed that the year is in this century (20yy).

The year part must be at the beginning or at the end (not in the middle).

The day and month parts may have 1 or 2 digits and may appear at any position.

let
    fnConvertDate = (DateString as text, Sequence as text) as date =>
    let
        Year1 = if Text.PositionOfAny(Sequence,{"y","Y"}) = 0 
                    then Text.Middle(DateString,0,4)
                    else Text.Middle(DateString,Text.PositionOfAny(DateString,{"/","-","."},Occurrence.Last)+1,4),

        Year2 = if Text.PositionOfAny(Year1,{"/","-","."}) = -1 
                    then Text.End("20"&Year1,4)
                    else Text.End("20"&Text.Start(Year1,Text.PositionOfAny(Year1,{"/","-","."})),4),

        Year = Number.From(Year2),

        Month1 = if Text.PositionOfAny(Sequence,{"m","M"}) = 0 
                    then Text.Middle(DateString,0,2)
                    else if Text.PositionOfAny(Sequence,{"m","M"}) = 1
                         then Text.Middle(DateString,Text.PositionOfAny(DateString,{"/","-","."})+1,2)
                         else Text.Middle(DateString,Text.PositionOfAny(DateString,{"/","-","."},Occurrence.Last)+1,2),

        Month = if Text.PositionOfAny(Month1,{"/","-","."}) = -1
                    then Number.From(Month1)
                    else Number.From(Text.Start(Month1,1)),

        Day1 = if Text.PositionOfAny(Sequence,{"d","D"}) = 0 
                    then Text.Middle(DateString,0,2)
                    else if Text.PositionOfAny(Sequence,{"d","D"}) = 1
                         then Text.Middle(DateString,Text.PositionOfAny(DateString,{"/","-","."})+1,2)
                         else Text.Middle(DateString,Text.PositionOfAny(DateString,{"/","-","."},Occurrence.Last)+1,2),

        Day = if Text.PositionOfAny(Day1,{"/","-","."}) = -1
                    then Number.From(Day1)
                    else Number.From(Text.Start(Day1,1))

    in
       #date(Year,Month,Day),

    Source = Excel.CurrentWorkbook(){[Name="DatesTable"]}[Content],
    Typed = Table.TransformColumnTypes(Source,{{"Dates", type text}, {"Sequence", type text}}),
    DateConverted = Table.AddColumn(Typed, "Date", each fnConvertDate([Dates],[Sequence]), type date)

in
    DateConverted

[ImkeF]

As far as I can see, Marcel's solution would work, if you manage to create a "Sequence"-column in the table that the "From-Folder" returns. So if you are able to describe the condition that identifies the format or merge with a lookup-table that contains that value, you could use this function.