CODEWITHSUNDEEP
phpserialization
Dejan Marjanović
Dejan Marjanović

Reputation: 19380

Standard serialization or custom?

can anyone recognize this format (if it is a standard format), or should I explode it manually to get arrays? Thanks.

{ coords : {lon : 7.41891, lat : 43.73253}, address : "", zipCode : "98000", city : "Monaco", sK : "Monaco", sQ : "852", fV : "", fZ : "98000 Monaco", fW : "- MC - Monaco: Monaco (98000)", gf : "31NDFzeHoxMGNORE11TnpNeU5UTT1jTnk0ME1UZzVNUT09", reflexId : "84167", areaLabel : "Monaco", jalon : 4}

YAML to VarExport gives this but then fails.

array (
  'coords' => 
  array (
    'lon' => 20.48406,
    'lat' => 44.80572,
  ),
)

Solution:

$data = preg_replace('#([\w]+) :#is', '"$1" :', $data);

Result:

array (
  'coords' => 
  {
     'lon' => 7.41891,
     'lat' => 43.73253,
  },
  'address' => '',
  'zipCode' => '98000',
  'city' => 'Monaco',
  'sK' => 'Monaco',
  'sQ' => '852',
  'fV' => '',
  'fZ' => '98000 Monaco',
  'fW' => '- MC - Monaco: Monaco (98000)',
  'gf' => '31NDFzeHoxMGNORE11TnpNeU5UTT1jTnk0ME1UZzVNUT09',
  'reflexId' => '84167',
  'areaLabel' => 'Monaco',
  'jalon' => 4,
)

Upvotes: 1

Views: 471

Answers (4)

jaywilliams
jaywilliams

Reputation: 311

After doing some research, it definitely looks like improperly formatted JSON. However, it's not too far off. In fact, with two regular expressions you could easily convert the string to proper JSON.

Find:
    /(\w+)\s+\:/
Replace:
    "$1" :

Find:
    /(})(\s+{)/
Replace:
    $1,$2

You can do it like this in PHP:

$good_json = preg_replace(array('/(\w+)\s+\:/i','/(})(\s+{)/'), array('"\1" :', '\1,\2'), $bad_json);
var_dump( json_decode($good_json) );

Upvotes: 0

Alin P.
Alin P.

Reputation: 44346

It's valid JavaScript code and almost JSON. For it to be valid JSON it would need to have the object property names enclosed in double quotes and have the array items separated by commas.

This:

$text = preg_replace('/(\r?\n){2}/', ',', trim($text));
$text = preg_replace('/([{,])\s?([a-z0-9_]+)\s?:/i', '$1"$2":',$text);
print_r(json_decode($text));

will work for your example.

But you should not use it as it will fail when there are , or { inside values!

The safest approach would be to construct a parser yourself.

Upvotes: 2

NikiC
NikiC

Reputation: 101936

Bad JSON. I would convert the hash keys to strings and then json_decode. You can do that with the PHP Tokenizer fairly simply.

function parse($code) {
    $result = '';
    $tokens = tokens_get_all('<?php ' . $code);
    array_shift($tokens); // drop <?php
    foreach ($tokens as $token) {
        if (!is_array($token)) {
            $result .= $token;
            continue;
        }

        if ($token[0] == T_STRING) {
            $token[1] = '\'' . addslashes($token[1]) . '\'';
        }

        $result .= $token[1];
    }

    return json_decode($result);
}

Upvotes: 0

Mikael Svenson
Mikael Svenson

Reputation: 39695

Looks like json serialization to me, so fairly standard. :)

Upvotes: 0

Related Questions