UTF-8 problems while reading CSV file with fgetcsv

Question

I try to read a CSV and echo the content. But the content displays the characters wrong.

Mäx Müstermänn -> Mäx Müstermänn

Encoding of the CSV file is UT

Answer 1

In my case the source file has windows-1250 encoding and iconv prints tons of notices about illegal characters in input string...

So this solution helped me a lot:

/**
 * getting CSV array with UTF-8 encoding
 *
 * @param   resource    &$handle
 * @param   integer     $length
 * @param   string      $separator
 *
 * @return  array|false
 */
private function fgetcsvUTF8(&$handle, $length, $separator = ';')
{
    if (($buffer = fgets($handle, $length)) !== false)
    {
        $buffer = $this->autoUTF($buffer);
        return str_getcsv($buffer, $separator);
    }
    return false;
}

/**
 * automatic convertion windows-1250 and iso-8859-2 info utf-8 string
 *
 * @param   string  $s
 *
 * @return  string
 */
private function autoUTF($s)
{
    // detect UTF-8
    if (preg_match('#[\x80-\x{1FF}\x{2000}-\x{3FFF}]#u', $s))
        return $s;

    // detect WINDOWS-1250
    if (preg_match('#[\x7F-\x9F\xBC]#', $s))
        return iconv('WINDOWS-1250', 'UTF-8', $s);

    // assume ISO-8859-2
    return iconv('ISO-8859-2', 'UTF-8', $s);
}

Response to @manvel's answer - use str_getcsv instead of explode - because of cases like this:

some;nice;value;"and;here;comes;combinated;value";and;some;others

explode will explode string into parts:

some
nice
value
"and
here
comes
combinated
value"
and
some
others

but str_getcsv will explode string into parts:

some
nice
value
and;here;comes;combinated;value
and
some
others

Answer 2

Try putting this into the top of your file (before any other output):

<?php

header('Content-Type: text/html; charset=UTF-8');

?>

Answer 3

The problem is that the function returns UTF-8 (it can check using mb_detect_encoding), but do not convert, and these characters takes as UTF-8. Тherefore, it's necessary to do the reverse-convert to initial encoding (Windows-1251 or CP1251) using iconv. But since by the fgetcsv returns an array, I suggest to write a custom function: [Sorry for my english]

function customfgetcsv(&$handle, $length, $separator = ';'){
    if (($buffer = fgets($handle, $length)) !== false) {
        return explode($separator, iconv("CP1251", "UTF-8", $buffer));
    }
    return false;
}

Answer 4

Now I got it working (after removing the header command). I think the problem was that the encoding of the php file was in ISO-8859-1. I set it to UTF-8 without BOM. I thought I already have done that, but perhaps I made an additional undo.

Furthermore, I used SET NAMES 'utf8' for the database. Now it is also correct in the database.

Answer 5

Encountered similar problem: parsing CSV file with special characters like é, è, ö etc ...

The following worked fine for me:

To represent the characters correctly on the html page, the header was needed :

header('Content-Type: text/html; charset=UTF-8');

In order to parse every character correctly, I used:

utf8_encode(fgets($file));

Dont forget to use in all following string operations the 'Multibyte String Functions', like:

mb_strtolower($value, 'UTF-8');

Answer 6

Try this:

<?php
$handle = fopen ("specialchars.csv","r");
echo '<table border="1"><tr><td>First name</td><td>Last name</td></tr><tr>';
while ($data = fgetcsv ($handle, 1000, ";")) {
        $data = array_map("utf8_encode", $data); //added
        $num = count ($data);
        for ($c=0; $c < $num; $c++) {
            // output data
            echo "<td>$data[$c]</td>";
        }
        echo "</tr><tr>";
}
?>