powershell script to change column position in text file - powershell

Question

Do you know how to use powershell to move column DocumentNo to last column ?
| D|Ref.Doc. |Row|DocumentNo |CoCd|Pstng Date
| W|5007534739| 1|65713191 |STCD|01/17/2016
| W|5007534739| 1|65713191 |STCD|01/17/2016
| W|5007534739| 1|65713191 |STCD|01/17/2016
Expected output
| D|Ref.Doc. |Row|CoCd|Pstng Date|DocumentNo
| W|5007534739| 1|STCD|01/17/2016|65713191
| W|5007534739| 1|STCD|01/17/2016|65713191
| W|5007534739| 1|STCD|01/17/2016|65713191
Here the command that I've tried
(get-content $file -ReadCount 0) |
foreach {
'{0}|{1}|{2}|{3}|{5}|{6}|{4}' -f $_.split('|')
} | Set-Content $file2
The code works properly, but in case DocumentNo containt a pipe as data, how to handle it ?

Answer 1

A delimeter is no good if it's used inside the data itself as it breaks the csv. I see that the data is fixed width (though the length and index may change from file to file), so I would take a different apporach.
Find the location for the DocumentNo-column. You can use regex for this (https://regex101.com/r/pH2oL9/1). I use [Regex]::Match() because it returns index (start position) and length (number of characters) of a match (the column).
Create a regex that finds the content at the same position and length on every line. It makes groups for "before", "content", "after" by counting "any character" until the start position I've decided, then the length of the column, then "until end of line". I use subexpressions $() in the regex to insert the index and length from step 2 since they might not be the same for every file.
Use -replace with the generated regex to modify the text on each line (since $text is an array). -replace finds the groups, and by using $1$3$2 I can say which order I want the groups inserted in the result. (https://regex101.com/r/vE6vO9/1)
Solution:
#Sample text
$text = Get-Content .\Test.txt
#Analyze header (find DocumentNo placement in fixed-width file) and create regex
$regex = [regex]::Match($text[0], '\|DocumentNo\s+') | ForEach-Object { "^(.{$($_.Index)})(.{$($_.Length)})(.*)$" }
#Modify text
$text -replace $regex, '$1$3$2' | Set-Content .\TestOut.txt
TestOut.txt
| D|Ref.Doc. |Row|CoCd|Pstng Date|DocumentNo
| W|5007534739| 1|STCD|01/17/2016|65713191
| W|5007534739| 1|STCD|01/17/2016|65713191
| W|5007534739| 1|STCD|01/17/2016|65713191
You might want to trim the trailing whitespaces. Use Trim() for this:
$text -replace $regex, '$1$3$2' | ForEach-Object { $_.Trim() } | Set-Content .\TestOut.txt

Answer 2

You could do the following:
$delimiter = "|"
$data = Get-Content "c:\tmp\test.csv";
$newCsv= ($data|Foreach-object { ($_ -split $delimiter)[#(0..3;5;6;4)] -join $delimiter})
# Set the new-ordered column content to the new-file
$newCsv|Set-Content C:\tmp\test2.csv
If you want all of this into one line you can do the following:
Get-Content "c:\tmp\test.csv"|Foreach-object { ($_ -split '|')[#(0..3;5;6;4)] -join '|'}|Set-Content C:\tmp\test2.csv
Note that you cannot use set-content if you have get-content as a part of the pipeline, because the file will be open and cannot replace it's content as a part of the stream.
You'd have to use either the first approach (read the content first, then pass it to the stream) or set the content into a different file.