Replacing part of a string in Powershell with regular expressions

Question

I have a problem:

I have an XML file which contains:

<colortable>
<color id="1" type="transparent"/>
<color id="2"/>
<color id="3" values="1.0"/>
<color id="4" type="rgb" values="0.0,0.0,0.0"/>
<color id="5" type="rgb" values="1.0,1.0,1.0"/>
</colortable>
<imagetable>
<imagedata id="1" source="E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.437248.1395746975.csfolha1v2SemMensagem_Tim.jpg">
</imagedata>
<imagedata id="2" source="E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.42189.1400584131.csfolha2v2fiscal_Tim.jpg">
</imagedata>
<imagedata id="3" source="E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.488328.1422006304.DT1_Image6_T.jpg">
</imagedata>
<imagedata id="4" source="E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.1262464.1427173896.csfolha3v2fiscal_Tim.jpg">
</imagedata>
<imagedata id="5" source="E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.54571.1400584131.csfolha0v2fiscal_Tim.jpg">
</imagedata>
</imagetable>

I want change the path from the one above to C:\images\

I'm trying to use this Powershell code:

while ($line = [Console]::In.ReadLine()) 
{ 
  switch -wildcard ($line) 
  { 
   '<imagedata*' {$line -replace '[A-Z]{1}:.+[r][.]([0-9]+[.]){2}', 'c:\images\'} 
   default {$line}
  }
}

I want this to, for every string starting with <imagedata, find the path (matching a regexp) and replace it with a new path.

This isn't working. How can I fix it?

`C:\images` instead of what ? Because last part seems to be name of some image file which I think you would want to keep as it is. And what is output of your current code ? — user2705585
– user2705585, Commented Mar 17, 2016 at 16:57
Works fine for me. Don't know what you want to keep but it replaces the path (and the "id"-part of the filename) with c:\images\ . The most disturbing thing in the script is [Console]::In.ReadLine()) which I really can't understand why you're using. — Frode F.
– Frode F., Commented Mar 17, 2016 at 19:11

Spencer D · Accepted Answer · 2016-03-17 17:26:49Z

1

I have not used PowerShell before, but I saw this under the RegEx tag, so I figured I would give it a look. I believe your problem comes from: :.+[r]. This is a greedy operator which could cause some issues for you. Perhaps try the following instead:

while ($line = [Console]::In.ReadLine()) 
{ 
  switch -wildcard ($line) 
  { 
   '<imagedata*' {$line -replace '[A-Za-z]:.+?r\.(\d+\.){2}', 'c:\images\'} 
   default {$line}
  }
}

Adding a question mark (?) makes the + lazy, instead of being greedy, which should then allow you to properly anything up to r.######.#######. (inclusive). I also swapped out your [.] with \. to indicate a little period mark, and I swapped [0-9] with \d (which is just a shorter way to write it). As someone pointed out the comments, though, are you sure you want to replace the r.######.#######. section?

If you end up wanting to keep the r.######.########. part of the filename, this should be accomplishable by using the RegEx positive lookahead operator instead of actually matching it. However, as a warning, some RegEx engine implementations throw a fit when the literal length of the lookahead/lookbehind is not specified/determinable. I am not sure how PowerShell handles varying length lookaheads, but here would be an implementation using that (assuming PowerShell supports it):

while ($line = [Console]::In.ReadLine()) 
{ 
  switch -wildcard ($line) 
  { 
   '<imagedata*' {$line -replace '[A-Za-z]:.+?(?=r\.(\d+\.){2})', 'c:\images\'} 
   default {$line}
  }
}

As an example, in your line that says: E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.437248.1395746975.csfolha1v2SemMensagem_Tim.jpg, rather than replacing: E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\r.437248.1395746975., the second code would, in theory, only replace: E:\xml2pdf_universal_physical_layer\tmp\dbres22C79BB2A484491458226919210\. Thus, it would preserve the whole filename. Again, this is dependent on PowerShell's support for lookahead and you may actually want to replace the first part of the filename, but I just wanted to throw that in here as an alternative, in case you do actually want to preserve the whole filename.

I hope that helps. Let me know if anything is unclear. You can read more about RegEx, and specifically lookahead and lookbehind, by clicking here (link to regular-expressions.info).

answered Mar 17, 2016 at 17:26

Spencer D

3,4862 gold badges31 silver badges45 bronze badges

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

2 Comments

TheMadTechnician Over a year ago

You may not know PowerShell but your solution is flawless. I just wanted to add a link to a RegEx101 explaination with the sample XML, and your solution in place. Also, if they wanted to keep the beginning if the file names, and had issues with using a lookahead they could use this for the replacement text: 'c:\images\$1'

Spencer D Over a year ago

Well, I'm glad to know it works, @TheMadTechnician lol Thanks for the additional link and suggestion about replacements. I only know RegEx because I spent a few years using it pretty heavily in VB.NET and PHP for scraping, so powershell specific things (suchas 'C:\images\$1') are pretty much lost on me. Glad to know my solution is correct, though.

Collectives™ on Stack Overflow

Replacing part of a string in Powershell with regular expressions

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related