split arguments with regex

Question

I'm using the following code to split incoming requests:

    /*
     *   Create global $GET from our cool URL's
     */
    preg_match_all('/[\/|\?|\&]([a-zA-Z\_0-9]+)([\/|=]([0-9]+)){0,1}/', $_SERVER['REQUEST_URI'], $parts, PREG_PATTERN_ORDER);
    if( count($parts[1]) >= 1 && count($parts[3]) >= 1)
        $GET = array_combine($parts[1], $parts[3]);

URL's like "/cat/2/post/345/answer/post" is divided into:

cat  => 2
post => 345
answer => 
post =>

is it possible to change this regex to allow for urls like: /post/345-title-goes-here/ and make it into:

post => 345-title-goes-here

becasue right now it only works if the argument is strictly 0-9 with no other characters. and even that was a real problem for me. any suggestions? :)

I've once again revisited this one, since the old one didn't seem to support single digit values. This is what I use now: #[/?&]([\w.()-]+)([/=](\d[\w.()-]+|\d))?# — Christoffer Bubach
– Christoffer Bubach, Commented May 18, 2015 at 2:22

Toto · Accepted Answer · 2012-01-20 16:26:56Z

1

I'd use a more concise one:

#[/?&](\w+)([/=](\d[\w-]+))?#

so the preg_match_all becomes:

preg_match_all('#[/?&](\w+)([/=](\d[\w-]+))?#', $_SERVER['REQUEST_URI'], $parts, PREG_PATTERN_ORDER);

Update according to comment:

#[/?&]([\w.()-]+)([/=](\d[\w.()-]+))?#

You can add every character you want within the character class.

edited Jan 20, 2012 at 16:26

answered Jan 19, 2012 at 19:37

Toto

91.7k63 gold badges97 silver badges135 bronze badges

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

6 Comments

Christoffer Bubach Over a year ago

neat short version , but it makes a url like /registrera/123-hejsan/doh/eller into [registrera] => 123-hejsan , [doh] => eller instead of [doh] => , [eller] => what i mean is that for it to count as value, i want it to start with 0-9.

Christoffer Bubach Over a year ago

Nice! It sure is short and tidy -evident that you know you'r regex well! Would it be totally ruined if I asked you for supporting signs . - _ ( ) anywhere in both key & value? Only real diffrence I want for values is that it should start with a digit, otherwise, the more URL friendly characters that can be supported on both key&value the better.

Toto Over a year ago

@ChristofferBubach: Of course, you can add any character you want. See my edit.

Christoffer Bubach Over a year ago

Hi, this was a while ago and I've been happily using this version on my sites: '#[/?&]([\w-.]+)([/=](\d[\w-.]+))?#' (not sure anymore how this one differs from your updated one above) problem is that this code along with every other version suggested here all fails if I get a URL with "/part/part/" which I discovered due to a relocation error. It should present this as two distinct keys, but only the first one is shown in the array. Any ideas how to solve this?

Christoffer Bubach Over a year ago

Seems like my issue here was that PHP arrays can't hold non-unique keys, so array_combine would simply ignore one of them.

|

Mike Buland · Accepted Answer · 2012-01-19 17:36:25Z

1

Sure, instead of:

'/[\/|\?|\&]([a-zA-Z\_0-9]+)([\/|=]([0-9]+)){0,1}/'

try:

'/[\/|\?|\&]([a-zA-Z\_0-9]+)([\/|=]([0-9]+[a-zA-Z_-]*)){0,1}/'

I just added the extra characters after the 0-9 in an optional (0 or more) set.

answered Jan 19, 2012 at 17:36

Mike Buland

7314 silver badges5 bronze badges

3 Comments

Christoffer Bubach Over a year ago

Doesn't this require that all parameters have a value? For example: /forum/5-general-talk/thread/45-hi-there/post/answer/ would result in all right until "post => answer" instead of "post =>" and "answer =>" ?

Christoffer Bubach Over a year ago

The regex need to somehow check that the following string consists of 0-9 then a "-" sign and then any other A-Z,a-z,- after that. Otherwise if it starts with a-z consider it a new parameter/variable.

Christoffer Bubach Over a year ago

Damn I spoke too soon, worked great! Got my hand on this from another buddy, seems to do even more, allows - in key-name for example. '/[\/|\?|\&]([a-zA-Z\-_0-9]+)([\/|=]([0-9]+\-[a-zA-Z()\.\-_0-9]+|[0-9]+)){0,1}/'

Collectives™ on Stack Overflow

split arguments with regex

2 Answers 2

6 Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

6 Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related