In PowerShell how do I use regex to validate input to include any character including backslashes?

Question

I would like to validate an input.

I want the input to be either blank, or allow for entries surrounded by single quotes separated by a comma, allowing for one, two, or more entries with any characters and any length between the single quotes. ex: 'one','two','three','\four','.dog','cat'

If I use this:

$string = Read-Host "Enter a string"
$pattern = "^$|^'\w+'(,'(\w+)')*$"
if ($string -match $pattern) {
    Write-Host "The string meets the pattern."
} else {
    Write-Host "The string DOES NOT meet the pattern."
}

For empty entry this works fine: ^$|

^'\w+'(,'(\w+)')*$ works for characters a-z,A-Z,0-9. But how do I allow for any character entry, including backslashes, periods and other special characters?

Thank you.

Ömer Sezer · Accepted Answer · 2023-12-15 22:40:24Z

3

Pls use $pattern = "^$|^'[^']*'(,'[^']*')*$".

[^']* matches any character except the single quote ('). It allows for any character between the single quotes, including backslashes, periods, and other special characters.

^$|^ => either an empty string or the following pattern.
'[^']*' => ': matches the opening single quote. ^: negates the character class, *: means zero or more occurrences. So, [^']* matches any character except the single quote ('). ': matches the closing single quote.
(,'[^']*')* => for a comma-separated list of single-quoted strings (zero or more times).
*$ => for the entire pattern to repeat zero or more times.

Sample Code:

$string = Read-Host "Enter a string"
$pattern = "^$|^'[^']*'(,'[^']*')*$"
if ($string -match $pattern) {
    Write-Host "The string meets the pattern."
} else {
    Write-Host "The string DOES NOT meet the pattern."
}

Sample Outputs:

Enter a string: 'one','two','three','\four','.dog','cat'
The string meets the pattern.

Enter a string:
The string meets the pattern.

Enter a string: '.dog','\cat'
The string meets the pattern.

edited Dec 15, 2023 at 22:40

answered Dec 15, 2023 at 22:27

Ömer Sezer

2,7327 gold badges17 silver badges19 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Santiago Squarzon Over a year ago

beat me to it with the same exact pattern haha, tho I think '[^']*' should be '[^']+' unless OP wants to allow '' as input. Might need feedback on his side on that.

HTWingNut Over a year ago

@SantiagoSquarzon - Good point, thank you. Yes, that is a fair consideration. It is best to not allow for '' as input.

mklement0 · Accepted Answer · 2023-12-16 03:28:55Z

To answer the question implied by the post's title:

It is the . regex metacharacter that represents any character - albeit by default excluding \n, i.e. a Unix-format newline, technically a LF character (LINE FEED, U+000A).
- To make . also match the latter, enable the single-line regex option: place (?s) at the start of the regex (in the simplest case, to apply to the entire expression).
Using a negated character class, such as [^'] to match anything but ', implicitly matches a newline too, even without setting the single-line option.
- Independently of whether . matches newlines or not, a run of characters - .* (if possibly empty) or .+ (if xnon-empty) - can be qualified with ? to make matching non-greedy, so that, in the (single-line) case at hand, .+? is an alternative to [^']+, i.e. matches only up to (but excluding) the closest ' char., if any.

As for a specific solution, to provide an alternative to Ömer Sezer's helpful answer:

The following regex uses a single capture group ((...)) to capture all tokens in the input string:
- ^(?:'([^']+)',?)*$ - explanation on regex101.com
- The above tolerates a trailing , - to avoid that, use:
  - ^(?:'([^']+)'(?:,(?!$))?)*$ - explanation on regex101.com
In combination with using [regex]::Match() in lieu of -match, this allows you to extract the individual tokens represented in the input string, as shown next:

$string = Read-Host 'Enter a string'

$match = [regex]::Match($string, "^(?:'([^']+)',?)*$")
if ($match.Success) {

  Write-Verbose -Verbose ('The string matches, ' + ('but is empty.', 'with the following tokens:')[[bool] $match.Value])
  $match.Groups[1].Captures.Value

} else {

  Write-Warning "The string does NOT match the pattern."

}

Note:

Accessing .Groups[1] on the [Match] instance returned by [regex]::Match() returns what the first (and only) capture group captured:
- .Captures accesses the collection of the multiple instances of the capture-group matches, themselves expressed as [Match] instances.
- .Value uses member-access enumeration to extract the matched text from each instance.
While the use of "..." is convenient here - because it allows use of ' without escaping - it is generally preferable to use '...' to represent regexes, i.e. verbatim string literals - see this answer.

Sample output, with 'foo','bar','baz' provided as verbatim input to the Read-Host command:

VERBOSE: The string matched, with the following tokens:
foo
bar
baz

Collectives™ on Stack Overflow

In PowerShell how do I use regex to validate input to include any character including backslashes?

2 Answers 2

2 Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related