regex to extract a word

Question

I'm trying to extract a single word from a line of text. As I understand it, Powershell regexes are almost the same as PCREs (and I have a way of easily testing PCREs). I have a file containing (amongst other things) something like...

ignore=thisline
username=symcbean
dontRead=thisEither

And I want to get the value associated with "username".

I know that the LHS of '=' will contain "username", optionally surrounded by whitespace, and the RHS will contain the value I am trying to extract (optionally surrounded by whitespace). The string I am looking for will match \w+, hence:

(?<=username=)\w+ 

works for the case without additional whitespace. But I can't seem to accommodate the optional white space. For brevity I've only shown the case of trying to handle a whitespace before the '=' below:

(?<=username\s*=)\w+   - doesn't match with or without additional space
(?<=username\W+)\w+    - doesn't match with or without additional space
(?<=username[\s=]*)\w+ - doesn't match with or without additional space

However in each case above, the group in the look-behind zero-width assertion (/username\s*=/, /username\W+/, /username[\s=]*/) matches the relevant part of the string.

I'm hoping to get a single value match (rather than array).

Answer 1

Meh, you could use regexes but then you would have two problems. This is how I would do it:

# Notice the extra spaces
$initialText = ' username = wombat  '
$userName = $intialText.Split('=')[1].Trim()

Here's how the key line works:

• The Split() method takes the string $initialText, and divides it into an array, eliminating the character passed to the split (treating it as a delimiter). So now, you have an array @(' username ',' wombat ').
• You then take the 1th (zero origin) element of the array ([1]). This is ' wombat '.
• You then call the Trim() method, which gets rid of all the whitespace at the beginning and the end of the string - so you now have 'wombat'...
• ...which you assign to $userName.

Split would still work you would just have to find the line that starts with username. Having said that, here is a regex method:

$initialText = ' username = wombat  '
$initialString -match '^.+=\W+(?<username>.+)\W+$'
$username = $matches.username

Or for an entire file:

From the prompt:

Get-Content C:\Path\To\Some\File.txt | %{if($_.trim().startswith('username')){$_ -match '^.+=\W?(?<username>.+)\W?$'; $username = $matches.username; $username}}

Or if you are doing it in a script:

$fileContents = Get-Content C:\Path\To\Some\File.txt
foreach($line in $fileContents){
  if($line.Trim().StartsWith('username')){
    $line -match '^.+=\W?(?<username>.+)\W?$'
    $userName = $matches.username
  }
}
$userName

Answer 2

This should do the trick if you are looking for multiple usernames in a single file. It will just put all values into an array of strings. The regular expressions pointed out should pull out what you want.

[regex]$username = "(?<=username\s*=\s*)\w+"
$usernames = @(Select-String -Path $file -Pattern $username -AllMatches | ForEach-Object {
    $_.Matches | ForEach-Object{
        $_.Value
    }
})

To explain a little bit of the Select-String commandlet, when you use the -AllMatches switch with it, it will return a collection of match objects. Inside those match objects are Matches, Groups, captures, etc. For this reason, you have to do the Foreach-Object { $_.Matches and then inside each matches object there is a value property hence | Foreach-Object { $_.Value

If it is only one username per file, you could just do this per file:

$text = get-content $file
[regex]$usernameReg = "(?<=username\s*=\s*)\w+"
$username = $usernameReg.Match($text).Value