Php json decode обработка ошибок

I just spent an hour going through all the possible solutions on this page. I took the liberty of collective all these possible solutions into one function, to make it quicker and easier to try and debug.

I hope it can be of use to someone else.

<?php
/**
 * Decontaminate text
 *
 * Primary sources:
 *  - https://stackoverflow.com/questions/17219916/json-decode-returns-json-error-syntax-but-online-formatter-says-the-json-is-ok
 *  - https://stackoverflow.com/questions/2348152/detect-bad-json-data-in-php-json-decode
 */
function decontaminate_text(
  $text,
  $remove_tags = true,
  $remove_line_breaks = true,
  $remove_BOM = true,
  $ensure_utf8_encoding = true,
  $ensure_quotes_are_properly_displayed = true,
  $decode_html_entities = true
){

  if ( '' != $text && is_string( $text ) ) {
    $text = preg_replace( '@<(script|style)[^>]*?>.*?</\1>@si', '', $text );
    $text = str_replace(']]>', ']]&gt;', $text);

    if( $remove_tags ){
      // Which tags to allow (none!)
      // $text = strip_tags($text, '<p>,<strong>,<span>,<a>');
      $text = strip_tags($text, '');
    }

    if( $remove_line_breaks ){
      $text = preg_replace('/[rnt ]+/', ' ', $text);
      $text = trim( $text );
    }

    if( $remove_BOM ){
      // Source: https://stackoverflow.com/a/31594983/1766219
      if( 0 === strpos( bin2hex( $text ), 'efbbbf' ) ){
        $text = substr( $text, 3 );
      }
    }

    if( $ensure_utf8_encoding ){

      // Check if UTF8-encoding
      if( utf8_encode( utf8_decode( $text ) ) != $text ){
        $text = mb_convert_encoding( $text, 'utf-8', 'utf-8' );
      }
    }

    if( $ensure_quotes_are_properly_displayed ){
      $text = str_replace('&quot;', '"', $text);
    }

    if( $decode_html_entities ){
      $text = html_entity_decode( $text );
    }

    /**
     * Other things to try
     * - the chr-function: https://stackoverflow.com/a/20845642/1766219
     * - stripslashes (THIS ONE BROKE MY JSON DECODING, AFTER IT STARTED WORKING, THOUGH): https://stackoverflow.com/a/28540745/1766219
     * - This (improved?) JSON-decoder didn't help me, but it sure looks fancy: https://stackoverflow.com/a/43694325/1766219
     */

  }
  return $text;
}

// Example use
$text = decontaminate_text( $text );

// $text = decontaminate_text( $text, false ); // Debug attempt 1
// $text = decontaminate_text( $text, false, false ); // Debug attempt 2
// $text = decontaminate_text( $text, false, false, false ); // Debug attempt 3

$decoded_text = json_decode( $text, true );
echo json_last_error_msg() . ' - ' . json_last_error();
?>

I’ll maintain it here: https://github.com/zethodderskov/decontaminate-text-in-php/blob/master/decontaminate-text-preparing-it-for-json-decode.php

object(stdClass)#1 (5) {
    ["a"] => int(1)
    ["b"] => int(2)
    ["c"] => int(3)
    ["d"] => int(4)
    ["e"] => int(5)
}

array(5) {
    ["a"] => int(1)
    ["b"] => int(2)
    ["c"] => int(3)
    ["d"] => int(4)
    ["e"] => int(5)
}

array(1) {
  [1]=>
  array(2) {
    ["English"]=>
    array(2) {
      [0]=>
      string(3) "One"
      [1]=>
      string(7) "January"
    }
    ["French"]=>
    array(2) {
      [0]=>
      string(3) "Une"
      [1]=>
      string(7) "Janvier"
    }
  }
}
Last error: JSON_ERROR_NONE

NULL
Last error: JSON_ERROR_DEPTH

object(stdClass)#1 (1) {
  ["number"]=>
  float(1.2345678901235E+19)
}
object(stdClass)#1 (1) {
  ["number"]=>
  string(20) "12345678901234567890"
}

Вернуться к: JSON Функции

One of the new features coming to PHP 7.3 is better error handling for json_encode() and json_decode(). The RFC was unanimously accepted by a 23 to 0 vote. Let’s take a look at how we handle JSON errors in <= PHP 7.2, and the new improvements coming in PHP 7.3.

Background

In the current stable, PHP v7.2, if you want to determine if JSON is invalid, you have to use the json_last_error() function to verify:

>>> json_decode("{");



  


                       
  



=> null
>>> json_last_error();
=> 4
>>> json_last_error() === JSON_ERROR_NONE
=> false
>>> json_last_error_msg()
=> "Syntax error"

For example, in the Laravel IlluminateEncryptionEncrypter class here’s a check to make sure calling json_encode() doesn’t result in an error:

// Once we get the encrypted value we'll go ahead and base64_encode the input
// vector and create the MAC for the encrypted value so we can then verify



  


                       
  



// its authenticity. Then, we'll JSON the data into the "payload" array.
$json = json_encode(compact('iv', 'value', 'mac'));
if (json_last_error() !== JSON_ERROR_NONE) {
    throw new EncryptException('Could not encrypt the data.');
}
return base64_encode($json);

We can at least determine if JSON encoding/decoding had an error, but it’s a little clunky compared to throwing an exception, which packages the error code and message neatly together.

Although you have to opt-in, v7.3 has an excellent way of allowing you to catch and handle JSON exceptions—let’s check out the new flag we can use!

The Throw on Error Flag in PHP 7.3

With the new option flag JSON_THROW_ON_ERROR it’s possible to rewrite this block of code with a try/catch. Maybe some something like the following:

use JsonException;



  


                       
  



try {
    $json = json_encode(compact('iv', 'value', 'mac'), JSON_THROW_ON_ERROR);
    return base64_encode($json);
} catch (JsonException $e) {
    throw new EncryptException('Could not encrypt the data.', 0, $e);
}

I think this new style is especially useful for userland code when you receive some JSON data and instead of digging around for json_last_error() and the matching flag, your JSON encoding and decoding can take advantage of the error handler.

The json_decode() function has a few more arguments, and will look something like the following in PHP 7.3 if you want to take advantage of error handling:

use JsonException;
try {



  


                       
  



    return json_decode($jsonString, $assoc = true, $depth = 512, JSON_THROW_ON_ERROR);
} catch (JsonException $e) {
    // Handle the JSON Exception
}
// Or even just let it bubble up...
/**
 * Decode a JSON string into an array
 *
 * @return array
 * @throws JsonException



  


                       
  



 */
function decode($jsonString) {
    return json_decode($jsonString, $assoc = true, $depth = 512, JSON_THROW_ON_ERROR);



  


                       
  



}

Getting the Error Code and Message

Previously you retrieve the JSON error code and message using the following functions:

// Error code
json_last_error();
// Human-friendly message
json_last_error_msg();

If you use the new JSON_THROW_ON_ERROR flag, here’s how you’ll get the code and message:

try {
    return json_decode($jsonString, $assoc = true, $depth = 512, JSON_THROW_ON_ERROR);
} catch (JsonException $e) {
    $e->getMessage(); // like json_last_error_msg()
    $e->getCode(); // like json_last_error()
}

See the base Exception class for more API details, the JsonException exception is a subclass of Exception.

JSON’s Default Behavior in PHP 7.3

When upgrading to PHP 7.3, your code will be backward compatible on day one and continue to work as expected.

PHP’s default json_encode|decode() behavior hasn’t changed, the throw on error RFC adds a new option and exception class.

Learn More

You can learn all about the most prominent features coming to PHP 7.3 in our post announcing the PHP 7.3 Alpha 1 release last week. Also, read through the RFC for JSON_THROW_ON_ERROR for full details on the proposal coming to PHP 7.3.

Unfortunately, the error handling in the earlier versions of PHP is non-existent. It wasn’t until PHP version 5.3.0 that the function json_last_error was introduced that. Before that, json_encode would return a NULL value or a FALSE value and you’d have to use an external JSON validation tool to try and figure out what was going wrong.

For example, if your data had special (non-UTF8) characters, the json_encode function would often return a NULL value. If the depth of your PHP array was too “deep”, then it would return a FALSE value. The json_decode function was also pretty similar in the way that it returned NULL whenever it received malformed / incorrect JSON strings.

Prior to PHP 5.3.0, JSON error handling in PHP looked a little like this (throwing Exceptions):

$val = json_encode($myArr);
if($val === false || is_null($val)){
    throw new Exception('Could not encode JSON');
}

or, with json_decode:

$val = json_decode($structure, true);
if(!is_array($val)){
    throw new Exception('Could not decode JSON');
}

Thankfully, in PHP 5.3.0, the function json_last_error was introduced. If an error has occurred, it will return one of the following constants:

• JSON_ERROR_NONE: No error has occured.
• JSON_ERROR_DEPTH: Max depth exceeded.
• JSON_ERROR_STATE_MISMATCH: Invalid JSON.
• JSON_ERROR_CTRL_CHAR: Bad character was found.

In PHP 5.3.3, another constant was added:

• JSON_ERROR_UTF8: Bad UTF8 character was found. Incorrect encoding.

Then, in PHP 5.5, they added three more constants:

• JSON_ERROR_RECURSION: Recursion detected.
• JSON_ERROR_INF_OR_NAN: One or more NAN or INF values in the value to be encoded.
• JSON_ERROR_UNSUPPORTED_TYPE: An unsupported type was found.

Using these constants, we can carry out our error handling like so:

<?php

//Attempt to decode JSON.
$decoded = json_decode($arr);

//Backwards compatability.
if(!function_exists('json_last_error')){
    if($decoded === false || $decoded === null){
        throw new Exception('Could not decode JSON!');
    }
} else{
    
    //Get the last JSON error.
    $jsonError = json_last_error();
    
    //In some cases, this will happen.
    if(is_null($decoded) && $jsonError == JSON_ERROR_NONE){
        throw new Exception('Could not decode JSON!');
    }
    
    //If an error exists.
    if($jsonError != JSON_ERROR_NONE){
        $error = 'Could not decode JSON! ';
        
        //Use a switch statement to figure out the exact error.
        switch($jsonError){
            case JSON_ERROR_DEPTH:
                $error .= 'Maximum depth exceeded!';
            break;
            case JSON_ERROR_STATE_MISMATCH:
                $error .= 'Underflow or the modes mismatch!';
            break;
            case JSON_ERROR_CTRL_CHAR:
                $error .= 'Unexpected control character found';
            break;
            case JSON_ERROR_SYNTAX:
                $error .= 'Malformed JSON';
            break;
            case JSON_ERROR_UTF8:
                 $error .= 'Malformed UTF-8 characters found!';
            break;
            default:
                $error .= 'Unknown error!';
            break;
        }
        throw new Exception($error);
    }
}

Unfortunately, this isn’t the type of code that you should be repeating throughout your script. If you stay true to the DRY (Don’t Repeat Yourself) principle (which you should), then you will wrap this kind of error checking into a function or a class method.

Personally, I would prefer it if json_last_error returned a string-based error message (like the internal XML library does). That way, you wouldn’t need to implement an ugly-looking switch statement.

Still, at least it is far more informative than it used to be!