Why does the PHP json_encode function convert UTF-8 strings to hexadecimal entities?

Question

I have a PHP script that deals with a wide variety of languages. Unfortunately, whenever I try to use json_encode, any Unicode output is converted to hexadecimal entities. Is this the expected behavior? Is there any way to convert the output to UTF-8 characters?

Here's an example of what I'm seeing:

INPUT

echo $text;

OUTPUT

База данни грешка.

INPUT

json_encode($text);

OUTPUT

"\u0411\u0430\u0437\u0430 \u0434\u0430\u043d\u043d\u0438 \u0433\u0440\u0435\u0448\u043a\u0430."

score 446 · Accepted Answer · edited Aug 02 '20 at 22:06

446

Since PHP/5.4.0, there is an option called JSON_UNESCAPED_UNICODE. Check it out:

https://php.net/function.json-encode

Therefore you should try:

json_encode( $text, JSON_UNESCAPED_UNICODE );

edited Aug 02 '20 at 22:06

Zombo

1
55
342
375

answered May 11 '13 at 14:46

John Severinson

4,755
1
14
6

3

Aha. Thanks! I should have read the documentation more carefully. Thanks. – David Jones May 11 '13 at 14:47
3

JSON_UNESCAPED_UNICODE was introduced in PHP 5.4.0, and is unavailable in earlier versions. When using it in earlier versions you will get this error: "Warning: json_encode() expects parameter 2 to be long, string given in ...". See CertaiN's answer below for 5.3 solution. – Octavian Naicu Oct 01 '14 at 10:03
This also works with Danish letters Æ,æ,Ø,ø,Å,å Thank you! – ymerdrengene Jan 22 '16 at 11:23
Fantastic, this was the answer I was looking for! – randomizer Dec 22 '16 at 16:33
2

You just saved my life. THANKS. – Jon Zangitu Jan 10 '18 at 09:06
This is amazing. I have been looking for days and trying out different functions but none of them helped me except this one. Thanks. – Tomb_Raider_Legend Feb 26 '18 at 20:36

mpyw · Answer 2 · 2016-02-13T08:01:19.407

61

JSON_UNESCAPED_UNICODE is available on PHP Version 5.4 or later.
The following code is for Version 5.3.

UPDATED

html_entity_decode is a bit more efficient than pack + mb_convert_encoding.
(*SKIP)(*FAIL) skips backslashes itself and specified characters by JSON_HEX_* flags.

function raw_json_encode($input, $flags = 0) {
    $fails = implode('|', array_filter(array(
        '\\\\',
        $flags & JSON_HEX_TAG ? 'u003[CE]' : '',
        $flags & JSON_HEX_AMP ? 'u0026' : '',
        $flags & JSON_HEX_APOS ? 'u0027' : '',
        $flags & JSON_HEX_QUOT ? 'u0022' : '',
    )));
    $pattern = "/\\\\(?:(?:$fails)(*SKIP)(*FAIL)|u([0-9a-fA-F]{4}))/";
    $callback = function ($m) {
        return html_entity_decode("&#x$m[1];", ENT_QUOTES, 'UTF-8');
    };
    return preg_replace_callback($pattern, $callback, json_encode($input, $flags));
}

edited Feb 13 '16 at 08:01

answered May 11 '13 at 15:26

mpyw

5,331
4
29
35

1

Shouldn't the \u be \U i.e. uppercase? – malhal Jan 30 '14 at 20:01
4

Nice solution for PHP < 5.4 ;) – qdev Feb 13 '14 at 10:47
I was looking for 3 days to find this solution for Version 5.3 as my host didn't upgrade to 5.4. To me you are a Life saver and for being so complete I would rather mark this as accepted answer! – Laci Jun 07 '14 at 08:36
Fixed bug when string contains `\\ `. Newer version grabs `\\ ` on higher priority than `\u`. – mpyw Feb 13 '16 at 06:52
This should be added in the php library. Good Job. – Beraki Dec 19 '16 at 10:30

score 13 · Answer 3 · answered Sep 25 '18 at 17:11

13

You like to set charset and unescaped unicode

 header('Content-Type: application/json;charset=utf-8');  
 json_encode($data,JSON_UNESCAPED_UNICODE|JSON_PRETTY_PRINT);

answered Sep 25 '18 at 17:11

Adrian Romero

459
6
12

score 6 · Answer 4 · answered May 03 '18 at 21:45

6

One solution is to first encode data and then decode it in the same file:

$string =json_encode($input, JSON_UNESCAPED_UNICODE) ; 
echo $decoded = html_entity_decode( $string );

answered May 03 '18 at 21:45

Steffo Dimfelt

840
12
10

score 2 · Answer 5 · answered Apr 29 '19 at 12:56

Here is my combined solution for various PHP versions.

In my company we are working with different servers with various PHP versions, so I had to find solution working for all.

$phpVersion = substr(phpversion(), 0, 3)*1;

if($phpVersion >= 5.4) {
  $encodedValue = json_encode($value, JSON_UNESCAPED_UNICODE);
} else {
  $encodedValue = preg_replace('/\\\\u([a-f0-9]{4})/e', "iconv('UCS-4LE','UTF-8',pack('V', hexdec('U$1')))", json_encode($value));
}

Credits should go to Marco Gasi & abu. The solution for PHP >= 5.4 is provided in the json_encode docs.

score 1 · Answer 6 · answered Apr 02 '19 at 17:30

1

json_encode($text, JSON_UNESCAPED_UNICODE|JSON_UNESCAPED_SLASHES);

answered Apr 02 '19 at 17:30

Hoàng Vũ Tgtt

1,597
20
7

score -2 · Answer 7 · answered Jan 18 '19 at 17:08

The raw_json_encode() function above did not solve me the problem (for some reason, the callback function raised an error on my PHP 5.2.5 server).

But this other solution did actually work.

https://www.experts-exchange.com/questions/28628085/json-encode-fails-with-special-characters.html

Credits should go to Marco Gasi. I just call his function instead of calling json_encode():

function jsonRemoveUnicodeSequences( $json_struct )
{ 
    return preg_replace( "/\\\\u([a-f0-9]{4})/e", "iconv('UCS-4LE','UTF-8',pack('V', hexdec('U$1')))", json_encode( $json_struct ) );
}

score -6 · Answer 8 · edited May 23 '17 at 12:34

-6

Is this the expected behavior?

the json_encode() only works with UTF-8 encoded data.

maybe you can get an answer to convert it here: cyrillic-characters-in-phps-json-encode

edited May 23 '17 at 12:34

Community

1
1

answered May 11 '13 at 15:33

JoakimH

39
5

Why does the PHP json_encode function convert UTF-8 strings to hexadecimal entities?

8 Answers8

Linked

Related