Powershell - Regular Expression Multiple Matches

Frankstar picture Frankstar · Jul 19, 2014 · Viewed 31.7k times · Source

Maybe my reasoning is faulty, but I can't get this working.

Here's my regex: (Device\s#\d(\n.*)*?(?=\n\s*Device\s#|\Z))

Try it: http://regex101.com/r/jQ6uC8/6

$getdevice is the input string. I'm getting this string from the Stream/Output from a command line tool.

$dstate = $getdevice |
     select-string -pattern '(Device\s#\d(\n.*)*?(?=\n\s*SSD\s+|\Z))' -AllMatches |
     % { $_ -match '(Device\s#\d(\n.*)*?(?=\n\s*SSD\s+|\Z))' > $null; $matches[0] }
Write-Host $dstate

Output:

Device #0 Device #1 Device #2 Device #3 Device #4

Same output for the $matches[1], $matches[2] is empty.

Is there a way I can get all matches, like on regex101.com? I'm trying to split the Output/String into separate variables (one for Device0, one for Device1, Device2, and so on).

Update: Here's the Output from the command line tool: http://pastebin.com/BaywGtFE

Answer

Matt picture Matt · Jul 19, 2014

I used your sample data in a here-string for my testing. This should work although it can depend on where your sample data comes from.

Using powershell 3.0 I have the following

$getdevice | 
    select-string -pattern '(?smi)(Device\s#\d+?(.*?)*?(?=Device\s#|\Z))' -AllMatches | 
    ForEach-Object {$_.Matches} | 
    ForEach-Object {$_.Value}

or if your PowerShell Verison supports it...

($getdevice | select-string -pattern '(?smi)(Device\s#\d+?(.*?)*?(?=Device\s#|\Z))' -AllMatches).Matches.Value

Which returns 4 objects with their device id's. I don't know if you wanted those or not but the regex can be modified with lookarounds if you don't need those. I updated the regex to account for device id with more that one digit as well in case that happens.

The modifiers that I used

  1. s modifier: single line. Dot matches newline characters
  2. m modifier: multi-line. Causes ^ and $ to match the begin/end of each line (not only begin/end of string)
  3. i modifier: insensitive. Case insensitive match (ignores case of [a-zA-Z])

Another regex pattern thats works in this way that is shorter

'(?smi)(Device\s#).*?(?=Device\s#|\Z)'