I am searching for following regex pattern in a file "[a-zA-Z0-9]+[^"]+[a-zA-Z0-9]+" This pattern contains all text including spaces. I am trying to Replace spaces(\s) to underscore(_) for all the entries in file that matches above pattern
(Get-Content .\temp.log -raw) -replace '("[a-zA-Z0-9]+[^"]+[a-zA-Z0-9]+")', {$($1.replace(' ','_'))}
Let me know what i am missing here.
When you use a script block ({ ... }) as the substitution operand of the -replace operator (doing so is only supported in PowerShell (Core) 7), you must refer to the match at hand via the automatic $_ variable:
(Get-Content .\temp.log -raw) -replace '("[a-zA-Z0-9]+[^"]+[a-zA-Z0-9]+")',
{ $_.Groups[1].Value -replace ' ', '_' }
Note:
.Groups[1].Value extracts the text that the first capture group ((...)) in your regex matched, because $_ contains a [System.Text.RegularExpressions.Match] instance.
$_.ValueAs for what you tried:
A symbol such as $1 to refer to what the 1st capture group captured only works if you pass a string as the substitution operand; even though it looks like a PowerShell construct, it is actually interpreted by the .NET regex engine that underlies PowerShell's -replace operand.
As an aside: The fact that $ has both special meaning to the .NET regex engine - both in regexes (which are always strings) and substitution operands (if passed as strings) - and PowerShell is the reason that it is better to use verbatim (single-quoted) strings, '...' rather than expandable (interpolating, double-quoted strings, "...").
I think a simple match-THEN-replace should work for the best:
# Here keeping the regex separated for better reading.
$RegexPattern='("[a-zA-Z0-9]+[^"]+[a-zA-Z0-9]+")'
# This will replace spaces with underscores in every line matching the Regex.
(Get-Content .\set.txt ) -match $RegexPattern -replace ' ', '_'
@(Get-Content ...) rather than(Get-Content ...))
'a' -replace '(a)',('$1' -replace 'a','b')