PHP-正则表达式删除引号并添加大括号？

Question

Well, I hate to admit it but I have a hard time with REGEX, I could never find a decent tutorial on how expressions should be set up.

So say I have something like this

context['something']

and I want to change all occurrences to

context[something] 

Then I have

' . $var . ' 

and I want to change all occurrences to

{var} 

This is the current concept but I am having trouble with the regex part. I am using str_replace but with language changes I don't think it would stable enough that way.

Here is my attempt.

$codes = array (
         '/(\' \. \$)(.+)( \. \')/',
        '/(\[\')(.+)(\'\])/'
);
$html = array (
        '{\\2}',
        '[\\2]',
);
$data = preg_replace($codes, $html, $data); 

It works until you get allot of them in a file and then it goes bad.

This is the current setup

// these are temp need a better replace system
$data = str_replace("' . $", "{", $data);
$data = str_replace(" . '", "}", $data);
$data = str_replace("<?php", "", $data);
$data = str_replace("?>", "", $data);
$data = str_replace('context[\'forum_name\']', 'context[forum_name]', $data); 

Just need a proper way to comment these so they can be converted back later on during save.

Can anyone please help?

Thanks :)

Answer 1

Make your matches non-greedy .

Change

.+

to

.+?

It works!!

Explanation:

The quantifier + is greedy by default.

Consider the regex \\[(.+)\\] which tries to match and capture everything between the [ and ] . On the input [foo] it works fine and foo is captured. But on the input [foo] and [bar] it'll capture foo] and [bar !!! This is the greedy behavior in action which makes + to consume as much as it can. By making the match non-greedy \\[(.+?)\\] we tell + to consume as less as it can but still trying to match. So in this case with the same input it'll capture foo .

Some tips:

There is no real need of the 1st and the 3rd group in your regex: '/(\\' \\. \\$)(.+)( \\. \\')/' . So you can drop the 1st and 3rd pair of (...) . If you just want to group use (?:..) rather than (..) which is used for grouping + capturing.

The use of \\\\n in the replacement part is discouraged. It should be used only as back-reference in the regex. To use capture groups in the replacement part use $n .

Program with above changes

Answer 2

我发现该工具在学习正则表达式时非常有用。

Answer 3

Here's one:

preg_replace("/context\['([^\']+)'\]/", "context[$1]", "context['bob']");

instead of using a non-greedy match, we just look for anything that's not a quotation mark.

And the other:

preg_replace("/' . [$]([^ ]+) . '/", '{$1}', '\' . $foo . \'');

This one's a bit trickier. We have to trap the $ in a set so it's not treated as the end-of-string character. And the {$1} has to be in single quotes so PHP doesn't interpret it + the braces as a variable.