Adding left padding using regex in Powershell

Question

I have a bunch of files, looking like this:

Prop.houses-2119-1
Prop.houses-2119-03

this should be renamed like this:

Prop<4-digits>house<2-digits>.house
Example:

Prop2119house01.house
Prop2119house03.house

So at the second block of numbers I have to add zeros if there is only one digit.

My code until now looks like this:

$propertyFiles = Get-ChildItem . *.property -rec
foreach ($file in $propertyFiles) {
(Get-Content $file) `
    -replace '(Prop.houses-)(\d+)(-)(\d+)', 'Prop$2house$4.house' `
    -replace 'second regex', 'second replacement' |
  Out-File $file
}

It is working fine, but I don't know how to add padLeft(2,'0') to the second block of numbers. It shows me the padLeft as a string.

The question is "I don't know how to add padLeft(2,'0') to the second block of numbers". The Use a function in Powershell replace explains how it is done. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Dec 21, 2023 at 14:38

Santiago Squarzon · Accepted Answer · 2023-12-21 14:58:21Z

You could use -match instead of -replace in this case, then using $Matches you can apply the zero left padding without issues:

$strings = @(
    'Prop.houses-2119-1'
    'Prop.houses-2119-03'
)

$strings | ForEach-Object {
    if ($_ -match '^(\D+)\.(\D+?)s?-(\d{4})-(\d{1,})$') {
        '{0}{1}{2}{3}.{2}' -f $Matches[1], $Matches[3], $Matches[2], $Matches[4].PadLeft(2, '0')
    }
}

In your code, applying this logic it would be:

foreach ($file in Get-ChildItem . *.property -rec) {
    (Get-Content $file) | ForEach-Object {
        # if the line matches the pattern
        if ($_ -match '^(\D+)\.(\D+?)s?-(\d{4})-(\d{1,})$') {
            # output the updated string
            '{0}{1}{2}{3}.{2}' -f $Matches[1], $Matches[3], $Matches[2], $Matches[4].PadLeft(2, '0')
            # and go next line
            return
        }

        # otherwise, output the line as is
        $_
    } | Out-File $file
}

mklement0 · Accepted Answer · 2024-01-02 23:11:53Z

The potential duplicate indeed shows the technique needed to implement your intent:

You want to perform a dynamic operation on a part of the matched string, based on its specific value, so as to be able to use a (conditionally) transformed version of it in the output string produced by a -replace-based string-substitution operation.
This can not be done with a string as the substitution operand,^[1] and instead requires a callback mechanism; in the context of PowerShell, it requires a script block ({ .. }), whose use with -replace is only supported in PowerShell (Core) 7+, however.
- In Windows PowerShell, you must call the underlying .NET API, [regex]::Replace(), directly, as shown in the linked post.
- The only substantial difference is in how the script block receives the match-information object (of type [Match]) that represents the matched string:
  - In PS 7+, with -replace, it is reflected in the automatic $_ variable (aka $PSItem).
  - With a direct [regex]::Replace() call, it is in the first positional argument, $args[0], via the automatic $args variable - though you're free to declare a parameter instead, as shown in the linked post (param($match)).

However, there's a twist to your requirements that to me warrants answering this question separately:

The need to not just perform a dynamic transformation, but to do so in the context of matching the input string as a whole and rearranging its - selectively transformed - substrings.

The immediate answer to your question - which is focused on an additional -replace operation, after having rearranged substrings - is the following:

# PS 7+ only
# Make sure that a single digit before '.house' is 0-padded to 2 digits.
# * $_ is a [System.Text.RegularExpressions.Match] instance with information
#   about the matched string.
# * $_.Value is the matched string itself, on which .PadLeft() can be called.
# -> 'Prop2119house01.house'
'Prop2119house1.house' -replace '(?<!\d)\d\b', { $_.Value.PadLeft(2, '0') }

Note:

This is essentially the same technique as in the linked post, but with the simplification of using a negative lookbehind assertion ((?<!...)) and a word-boundary assertion to (\b) so as to avoid having to use capture groups ((...)) - see this regex101.com page for an explanation of the regex and the option to experiment with it.

However, you can use this technique to make do with a single -replace operation, which combines rearranging substrings with selective dynamic transformations, based on matching and replacing the whole input string, with the substring of interest captured via capture groups ((...)):

# PS 7+ only

# Sample input lines.
$lines = 'Prop.houses-2119-1', 'Prop.houses-2119-03'

# * $_.Groups[1].Value is the text captured by the 1st capture group, and so on.
# -> 'Prop2119house01.house', 'Prop2119house03.house'
$lines -replace 'Prop\.houses-(\d+)-(\d+)', { 
  'Prop{0}house{1}.house' -f $_.Groups[1].Value, $_.Groups[2].Value.PadLeft(2, '0') 
}

Note the use of -f, the format operator, to assemble the final string; in the template string, {0} is a placeholder for the first RHS operand, and so on.

^{[1] For a detailed explanation, see the bottom section of this answer}

js2010 · Accepted Answer · 2023-12-28 22:52:29Z

0

If I understand this correctly...

$string = 'Prop.houses-2119-1'
$prop,$house,$year,$num = $string -split '\.|-'
$house = $house -replace 's$'
$num = [int]$num | % tostring 00
"$prop$year$house$num.$house"

Prop2119house01.house

answered Dec 28, 2023 at 22:52

js2010

28.3k6 gold badges82 silver badges88 bronze badges

Comments

nimizen · Accepted Answer · 2023-12-21 14:47:00Z

This can be achieved with basic string manipulation rather than a regex:

$files = Get-ChildItem "C:\Temp\.so" -File

foreach($f in $files)
{    
#splitting on the dashes isolates the numbers
    $splitOnDashes = $f.Name.split('-')

#splitting on the period isolates the Prop and Houses elements
    $propHouses = $splitOnDashes[0].Split('.')

#if the split element containing the one or two digit number is one number, pad a zero
    if($splitOnDashes[2].Length -eq 1)
    {
        $splitOnDashes[2] = "0$($splitOnDashes[2])"
    }
# bring all the elements together and rename the file, whilst also trimming the 's' from 'houses'.    
    Rename-Item $f.FullName "$($propHouses[0])$($splitOnDashes[1])$($propHouses[1].TrimEnd('s'))$($splitOnDashes[2]).$($propHouses[1].TrimEnd('s'))" -Force
}

Theo · Accepted Answer · 2023-12-21 16:31:30Z

-2

Yet another way is to use switch to loop over the lines in each file and update the lines that match a certain regex pattern:

$propertyFiles = Get-ChildItem . '*.property' -File -Recurse
foreach ($file in $propertyFiles) {
    $newContent = switch -Regex -File $file.FullName {
        # update strings that match this pattern
        '(Prop)\.?(house).*(\d{4})-?(\d+)$' { 
            '{0}{1}{2}{3:D2}.{2}' -f $Matches[1], $Matches[3], $Matches[2], [int]$Matches[4]
        }
        default { $_ }  # output unchanged when no match
    }
    $newContent | Set-Content -Path $file.FullName
}

answered Dec 21, 2023 at 16:31

Theo

61.9k8 gold badges28 silver badges48 bronze badges

Collectives™ on Stack Overflow

Adding left padding using regex in Powershell

5 Answers 5

Comments

Comments

Comments

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

5 Answers 5

Comments

Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related