Search code examples
regexparsingtokenizeregex-lookarounds

Regular expression to capture windows paths or filenames starting with X:\ **enclosed or not** by quotes to have spaces in it

I need to build a regular expression to capture one or more windows paths inside a text. It's for a syntax highlighter.

Imagine this text:

Hey, Bob!

I left you the report for tomorrow in D:\Files\Shares\report.pdf along
with the other reports.

There's also this pptx here D:\Files\Internal\source.pptx where you have
the original if you need to change anything.

Cheers!
Alice.

This one is easy to capture with /[a-zA-Z]:\\[^\s]*/mg. See it in regex101 here https://regex101.com/r/VcBV7M/1

Capturing sample 1

Nevertheless

when the path has spaces like here:

I left you the report for tomorrow in D:\Shared files\october report.pdf along
with the other reports.

then we run into problems: What is the path? D:\Shared or D:\Shared files\october or D:\Shared files\october report.pdf or D:\Shared files\october report.pdf along...

For a human it's simple to infer. For a computer it's impossible so I was thinking into forcing the users to use quotes or brackets to indicate the begin and end of the filename or path.

Question

How can I write a regex that given this:

Hey, Bob!

I left you the report for tomorrow in "D:\Shared files\october report.pdf" along
with the other reports [Don't forget to add your punctuation]. See
also D:\Multifiles\charlie.docx for more info.

There's also this pptx here [D:\Internal files\source for report.pptx] where you have
the original if you need to change "anything like the boss wants".

Cheers!
Alice.

captures this?

D:\Shared files\october report.pdf
D:\Multifiles\charlie.docx
D:\Internal files\source for report.pptx

but not

Don't forget to add your punctuation
anything like the boss wants

Non-working sample: https://regex101.com/r/RGVPz6/2

Solution

  • I found that lookahead and lookbehind helped.

    The expression is: (?<=\[)[a-zA-Z]:\\.*(?=])|(?<=")[a-zA-Z]:\\.*(?=")|([a-zA-Z]:\\[^\s]*)

    Here's the solution online-live-edit https://regex101.com/r/RGVPz6/3

    The key parts are:

    • In the first option (?<=\[) and (?=]) tell to only match if the string is exactly surrounded by [ and ].
    • In the second option (?<=") and (?=") tell to only match if the string is exactly surrounded by " and ".
    • The third option is the old one.

    See it working:

    Solution

    PD: Inspired here https://es.javascript.info/regexp-lookahead-lookbehind