正则表达式在函数体中找到某些单词

Question

I want to write a regular expression to find the word after $body->
That is userId, deviceId, deviceToken and deviceType in the following code sample:

 function deviceUpdate($body)
 {  
    $userId=trim($body -> userId);
    $deviceId= $body -> deviceId;
    $deviceToken=trim($body -> deviceToken);
    $deviceType=trim($body -> deviceType);
 }

There can be a single space or multiple spaces after body keyword and after ->

The regular expression I tried is:

(?<=body ->\s)\S+

But this does not work properly if there is no space between body and ->.

Edit : I used the the regular expression ~function\s*(.+?)(~ to find the name of function ( deviceupdate) and ~\$body\s*->\s*(\w+)~ to find all keywords after body -> operator. It works perfectly fine. Suppose, i have 2 functions :

function deviceUpdate($body)
 {  
    $userId=trim($body -> userId);
    $deviceId= $body -> deviceId;

 }

function postad($body)
 {

    $adtitle = $body -> adtitle;
    $addescription = $body -> addescription;

  }

How to extract the function 1 name that is deviceUpdate and corresponding keywords in that same function after body -> that is ( userId,deviceId) in one array and function 2 name postad and corresponding keywords in that same function after body -> that is ( adtitle,addescription) in second array?

Currently when i use preg_match_all and regular expression as stated above it returns everything it matches in a single array.

Answer 1

In your solution, you are using a positive lookbehind that, in PHP PCRE regex, cannot be of undefined length.

If you have cases where $body and -> are separated with zero or more spaces, you cannot rely on a fixed-width lookbehind. Use capturing and grab the submatch from the match object (to only match alphanumerics with underscore to avoid matching );, you need to replace \S with \w):

~\$body\s*->\s*(\w+)~

See the regex demo

See IDEONE demo:

$re = '~\$body\s*->\s*(\w+)~'; 
$str = " function deviceUpdate(\$body)
 {  
    \$userId=trim(\$body -> userId);
    \$deviceId= \$body -> deviceId;
    \$deviceToken=trim(\$body -> deviceToken);
    \$deviceType=trim(\$body -> deviceType);
 }"; 
preg_match($re, $str, $matches);
echo $matches[1]; // => userId

If you need all matches, use preg_match_all:

preg_match_all($re, $str, $matches);
print_r($matches[1]);

See another IDEONE demo

To address the edit:

You can use

(?:\bfunction\s+(\w+)\([^()]*\)\s*\{|(?!^\G))(?:(?!\bfunction\s+\w+\([^()]*\)\s+{|}\h*?
).)*?\$body\s*->\s*(\w+)

See the regex demo

In PHP, use it as

$re = '~(?:\bfunction\s+(\w+)\([^()]*\)\s*\{|(?!^\G))(?:(?!\bfunction\s+\w+\([^()]*\)\s+{|}\h*?
).)*?\$body\s*->\s*(\w+)~s'; 

See the PHP demo

The regex above uses a custom boundary specified with (?:\bfunction\s+(\w+)\([^()]*\)\s*\{|(?!^\G)) and a tempered greedy token (?:(?!\bfunction\s+\w+\([^()]*\)\s+{|}\h*? ).)* making sure a match is only found inside one function