How to Accurately Substring Chinese Text in PHP Across GBK and UTF-8

Key points to know:

In GBK encoding, a Chinese character occupies 2 bytes; in UTF‑8 it occupies 3 bytes.

The ord() function returns the ASCII value of the first character of a string.

Chinese characters have ASCII values greater than 0xA0.

The essential technique is to determine whether each character in the string is Chinese or English by checking if ord(substr($str,$start,1)) > 0xA0 . If true, it is a Chinese character; otherwise, it is an English character.

The following PHP function my_substr implements this logic, allowing you to specify the start position, length, and byte size (2 for GBK, 3 for UTF‑8):

<code><?php
/* param $str   The string to be truncated.
 * param $start Starting position, 0 for the first character.
 * param $length Number of characters to extract; if empty, extract to the end.
 * param $bite   Byte length of a Chinese character, default 2 for GBK, 3 for UTF‑8.
 */
function my_substr($str, $start, $length = "", $bite = 2) {
    $pos = 0; // byte position in the string
    // Calculate byte offset for the start position
    for ($i = 0; $i < $start; $i++) {
        if (ord(substr($str, $i, 1)) > 0xA0) {
            $pos += $bite; // Chinese character
        } else {
            $pos += 1; // English character
        }
    }

    if ($length == "") {
        return substr($str, $pos); // to the end
    } else {
        if ($length < 0) {
            $length = 0;
        }
        $string = "";
        for ($i = 1; $i <= $length; $i++) {
            if (ord(substr($str, $pos, 1)) > 0xA0) {
                $string .= substr($str, $pos, $bite);
                $pos += $bite;
            } else {
                $string .= substr($str, $pos, 1);
                $pos += 1;
            }
        }
        return $string;
    }
}
$str = "a这是一段中文";
echo my_substr($str, 0); // output whole string
echo "\n";
echo my_substr($str, 0, 1); // output 'a'
echo "\n";
echo my_substr($str, 1, 2); // output '这是一'
?>
</code>

Adjust the $bite parameter to 3 when working with UTF‑8 encoded strings.