Regex replace hexadecimal characters

Question

I am new to regex and unable to create a expression to resolve my problem.

I have a string like /abc/def%20hi%28hello%29test. I would like to replace the hexadecimal characters, that is, %20, %28 and %29 with -

I know all three of them have different meanings, but I am fine with replacing all with any single character, I mean - or even . or ' '(space).

So, I would like to get the output as /abc/def-hi-hello-test.

I thought to find % and replace three characters from %[a-zA-Z0-9][a-zA-Z0-9] with -, but doesn't work.

Can some please help me write a regex which just replaces those hexadecimal characters?

What do you mean when you say it "doesn't work"? Using your example, what output does it give? — Michelle, Aug 13 '13 at 14:54
Your regex will match all your examples... and actually more than you want since hex only goes `a-f`. — Smern, Aug 13 '13 at 14:56
We need to know [what you have tried](http://whathaveyoutried.com) in a more specific way - are you writing this in a program? or are you trying to use this in a text editor's search and replace dialogue? Regular Expressions are not language independent - you should consider writing a [Short, Self Contained, Compilable Example](http://sscce.org) so we can better diagnose your problem. — FrankieTheKneeMan, Aug 13 '13 at 14:56
Seems fine, you could potentially even change it to `%[A-Z0-9]{2}` or `%[A-Z\d]{2}` for easier reading (depending on Regex engine). — Chris, Aug 13 '13 at 14:57
@Chris You would need to remember the case-insensitive flag for that. — Bernhard Barker, Aug 13 '13 at 15:01
Your regex looks fine (except for including `G-Z`, which you were unlikely to have noticed since you just said it "doesn't work", which is really not specific enough). Please construct an [SSCCE](http://sscce.org/) in the language / tool of your choice, which should include an easily reproduceable test case which doesn't work. — Bernhard Barker, Aug 13 '13 at 15:03
@Dukeling Fair point, I'd assumed he'd grabbed a URL encoded something and he'd always have upper case letters (I'm now not sure why I left the entire range of letters in place). Derp. — Chris, Aug 13 '13 at 15:05
I'm under the impression the OP needs the global flag `g` (or equivalent). — Jerry, Aug 13 '13 at 15:10
You may need to complicate it quite a bit to be 100% correct - I would guess that `%%28` shouldn't become `%(`, because `%%` is an escaped `%`. — Bernhard Barker, Aug 13 '13 at 15:19

score 3 · Answer 1 · answered Aug 13 '13 at 15:01

3

Search for:

%[a-fA-F0-9][a-fA-F0-9]

replace with:

Example: http://regex101.com/r/qR6xW7

answered Aug 13 '13 at 15:01

mart1n

5,969
5
46
83

Making it shorter `(?i)%[a-f0-9]{2}` :) – HamZa Aug 13 '13 at 15:02
2

I don't think regex is the true issue the OP is having as his example would actually match. I believe it is the replace he is having problems with... which we can't really help him with without knowing the language he is using. – Smern Aug 13 '13 at 15:04
@HamZa Or even shorter by replacing `0-9` with `\d`. – Bernhard Barker Aug 13 '13 at 15:05
@HamZa, that is dependent on regex engine. – Smern Aug 13 '13 at 15:06
@Dukeling Well `\d` [also matches](http://stackoverflow.com/a/6479605) hindu, arabic or hebrew digits. – HamZa Aug 13 '13 at 15:08
@smerny You got the idea, setting the `i` modifier :) – HamZa Aug 13 '13 at 15:08
@HamZa That's news to me, though I'll probably continue to pretend they mean the same, at least in my pet projects, since I don't ever work with non-0-9 digits. – Bernhard Barker Aug 13 '13 at 15:18

score 1 · Answer 2 · answered Aug 13 '13 at 15:24

It sounds like what you really want to do is decode the URL encoding, and I will bet you that your host language has a facility for that built in. In PHP, that's url_decode.

After you've decoded the URL, then you can modify what it is you want to modify more easily.

Regex replace hexadecimal characters

2 Answers2