ansaurus

Question

How to get the previous word with regex?

Answer 1

+2 A:

It is not a good idea to parse paths with regular expressions - it's much better to use a specific library for that purpose. What language are you using?

Eli Bendersky 2010-01-27 08:04:10

I agree with you, but this isn't an answer.

Rubens Farias 2010-01-27 08:15:03

@Rubens, I'm trying to teach fishing, not just give a fish

Eli Bendersky 2010-01-27 08:16:57

I'm using C#, and I found a better solution for this, rather than using regex. You're right, it is better to use a specific library for that purpose, but first I also needed the regex to make some tests. Thank you anyway

Bogdan Craciun 2010-01-27 08:26:01

@Eli, you can do that as a comment;

Rubens Farias 2010-01-27 08:28:35

Answer 2

A:

You should find this pattern in string: ([\w\s.-]+)\\[\w\s.-]+$. The first group will containg word 'blue' in your case. Exact syntax of regex and accessing groups depends on your programming language.

Rorick 2010-01-27 08:04:10

`\d` is a subset of `\w`, so `\w\d` is redundant.

Amarghosh 2010-01-27 08:21:08

Yes, you're right.

Rorick 2010-01-27 08:23:31

Answer 3

+4 A:

(\w+)\W+\w+$

matches the second-to-last word, captures it into backreference no. 1, then matches one or more non-word characters, then the last word and then EOL.

If you don't really want actual words but path elements to match (even if they contain non-word characters), then

([^\\:]+)\\[^\\]+$

might do a better job.

Edit: Added a : to the "second-to-last-word group" so the regex can handle relative paths, too.

Tim Pietzcker 2010-01-27 08:05:05

Path elements contains non-word characters, e.g. 'R4-branch' contains '-'. So your regex won't match 'C:\Development.TFS\Examn\R4-branch\Web\OrganisationManager\themes\blue-sea\css'

Rorick 2010-01-27 08:14:49

You're right of course. The OP asked for words, though, not path elements (which he probably meant). Will edit.

Tim Pietzcker 2010-01-27 08:20:10

+1 for the updated regex.

Amarghosh 2010-01-27 08:23:43

Yes, this is nicer! It will even correctly work with spaces unlike mine =)

Rorick 2010-01-27 08:26:17

Thank you! Indeed, I actually wanted that path element, but your answer helped a lot :)

Bogdan Craciun 2010-01-27 08:27:30

But won't work with relative paths: 'C:blue\css' =)

Rorick 2010-01-27 08:30:38

Right again. Well, the more specific the requirements, the better the regex.

Tim Pietzcker 2010-01-27 08:55:04

Answer 4

A:

if (eregi('themes\\([a-z].+)\\css', $subject)) { # Successful match } else { # Match attempt failed }

A PHP example to get the word blue.

Marco 2010-01-27 08:16:17

Maybe you must try regex buddyit is a great tool to make regexes

Marco 2010-01-29 11:20:45

Answer 5

A:

can't you simply split the string using the character '\' then get the splitResult[splitResult.Count-1]?

you could always replace '\' by the path separator in your environment, for more consistent results

Sam Holder 2010-01-27 08:19:48

Answer 6

A:

Another option:

(\w+)\\\w+$

Alix Axel 2010-01-27 08:26:10

I posted this and deleted as the "words" in OP's path can include non `\w` characters like `-`,`.`,`space` etc.

Amarghosh 2010-01-27 08:43:09

ansaurus

tags:

views:

answers:

How to get the previous word with regex?

related questions