json returns null with special characters

Question

Today I came across this mistake:

When trying to encode an array for json, some values return null. I arrived at the solution before using the function json_encode() pass all strings to UTF-8 with htmlentities($string, UTF-8) but in a special case, a string had this character - (notice that it is different from a hyphen -), and returned me error. look at the tests I did:

 $string    = htmlentities('Av. 7 de setembro – Salvador', UTF-8); // String com o caracter que retorna erro
 $string2   = htmlentities('Av. 7 de setembro - Salvador', UTF-8); // String com o Hífen
 $array = array('string' => $string, 'string2' => $string2);
 echo json_encode($array);

 Output: {"string":null,"string2":"Av. 7 de setembro - Salvador"}

Is there another way to convert this character? I’m afraid there are other characters that htmlentities() will not solve.

Answer 1

Use the constant ENT_QUOTES to convert the quotes:

 $string    = htmlentities('Av. 7 de setembro – Salvador', ENT_QUOTES, "UTF-8"); // String com o caracter que retorna erro
 $string2   = htmlentities('Av. 7 de setembro - Salvador', ENT_QUOTES, "UTF-8"); // String com o Hífen
 $array = array('string' => $string, 'string2' => $string2);
 echo json_encode($array);

Example

Answer 2

I would use json_encode directly. But if you are having problems with special characters, you can try using the functions: utf8_encode or utf8_decode in place of htmlentities

Answer 3

Remove the htmlentities and directly use the json_encode:

<?php    
  header ('Content-type: text/html; charset=UTF-8');
  $string    = 'Av. 7 de setembro – Salvador'; // String com o caracter que retorna erro
  $string2   = 'Av. 7 de setembro - Salvador'; // String com o Hífen
  $array = array('string' => $string, 'string2' => $string2);
  echo json_encode($array);

Exit:

{"string":"Av. 7 de setembro \u2013 Salvador","string2":"Av. 7 de setembro - Salvador"}

\u2013: Unicode Character 'EN DASH' (U+2013)

Example: ideone

Answer 4

When I have this kind of problem I use the function "mb_convert_encoding"

  mb_convert_encoding($string, 'UTF-8', 'LATIN1');

See the function manual on http://php.net/manual/en/function.mb-convert-encoding.php