PHP: Split multibyte string (word) into separate characters - php

Question

Trying to split this string "主楼怎么走" into separate characters (I need an array) using mb_split with no luck... Any suggestions?
Thank you!

Answer 1

try a regular expression with 'u' option, for example
$chars = preg_split('//u', $string, -1, PREG_SPLIT_NO_EMPTY);

Answer 2

An ugly way to do it is:
mb_internal_encoding("UTF-8"); // this IS A MUST!! PHP has trouble with multibyte
// when no internal encoding is set!
$string = ".....";
$chars = array();
for ($i = 0; $i < mb_strlen($string); $i++ ) {
$chars[] = mb_substr($string, $i, 1); // only one char to go to the array
}
You should also try your way with mb_split with setting the internal_encoding before it.

Answer 3

You can use grapheme functions (PHP 5.3 or intl 1.0) and IntlBreakIterator (PHP 5.5 or intl 3.0). The following code shows the diffrence among intl and mbstring and PCRE functions.
// http://www.php.net/manual/function.grapheme-strlen.php
$string = "a\xCC\x8A" // 'LATIN SMALL LETTER A WITH RING ABOVE' (U+00E5)
."o\xCC\x88"; // 'LATIN SMALL LETTER O WITH DIAERESIS' (U+00F6)
$expected = ["a\xCC\x8A", "o\xCC\x88"];
$expected2 = ["a", "\xCC\x8A", "o", "\xCC\x88"];
var_dump(
$expected === str_to_array($string),
$expected === str_to_array2($string),
$expected2 === str_to_array3($string),
$expected2 === str_to_array4($string),
$expected2 === str_to_array5($string)
);
function str_to_array($string)
{
$length = grapheme_strlen($string);
$ret = [];
for ($i = 0; $i < $length; $i += 1) {
$ret[] = grapheme_substr($string, $i, 1);
}
return $ret;
}
function str_to_array2($string)
{
$it = IntlBreakIterator::createCharacterInstance('en_US');
$it->setText($string);
$ret = [];
$prev = 0;
foreach ($it as $pos) {
$char = substr($string, $prev, $pos - $prev);
if ('' !== $char) {
$ret[] = $char;
}
$prev = $pos;
}
return $ret;
}
function str_to_array3($string)
{
$it = IntlBreakIterator::createCodePointInstance();
$it->setText($string);
$ret = [];
$prev = 0;
foreach ($it as $pos) {
$char = substr($string, $prev, $pos - $prev);
if ('' !== $char) {
$ret[] = $char;
}
$prev = $pos;
}
return $ret;
}
function str_to_array4($string)
{
$length = mb_strlen($string, "UTF-8");
$ret = [];
for ($i = 0; $i < $length; $i += 1) {
$ret[] = mb_substr($string, $i, 1, "UTF-8");
}
return $ret;
}
function str_to_array5($string) {
return preg_split('//u', $string, -1, PREG_SPLIT_NO_EMPTY);
}
When working on production environment, you need to replace invalid byte sequence with the substitute character since almost all grapheme and mbstring functions can't handle invalid byte sequence. If you have an interest, see my past answer: https://stackoverflow.com/a/13695364/531320
If you don't take of perfomance, htmlspecialchars and htmlspecialchars_decode can be used. The merit of this way is supporting various encoding other than UTF-8.
function str_to_array6($string, $encoding = 'UTF-8')
{
$ret = [];
str_replace_callback($string, function($char, $index) use (&$ret) { $ret[] = $char; return ''; }, $encoding);
return $ret;
}
function str_replace_callback($string, $callable, $encoding = 'UTF-8')
{
$str_size = strlen($string);
$string = str_scrub($string, $encoding);
$ret = '';
$char = '';
$index = 0;
for ($pos = 0; $pos < $str_size; ++$pos) {
$char .= $string[$pos];
if (str_check_encoding($char, $encoding)) {
$ret .= $callable($char, $index);
$char = '';
++$index;
}
}
return $ret;
}
function str_check_encoding($string, $encoding = 'UTF-8')
{
$string = (string) $string;
return $string === htmlspecialchars_decode(htmlspecialchars($string, ENT_QUOTES, $encoding));
}
function str_scrub($string, $encoding = 'UTF-8')
{
return htmlspecialchars_decode(htmlspecialchars($string, ENT_SUBSTITUTE, $encoding));
}
If you want to learn the specification of UTF-8, the byte manipulation is the good way to practice.
function str_to_array6($string)
{
// REPLACEMENT CHARACTER (U+FFFD)
$substitute = "\xEF\xBF\xBD";
$size = strlen($string);
$ret = [];
for ($i = 0; $i < $size; $i += 1) {
if ($string[$i] <= "\x7F") {
$ret[] = $string[$i];
} elseif ("\xC2" <= $string[$i] && $string[$i] <= "\xDF") {
if (!isset($string[$i+1])) {
$ret[] = $substitute;
return $ret;
} elseif ($string[$i+1] < "\x80" || "\xBF" < $string[$i+1]) {
$ret[] = $substitute;
} else {
$ret[] = substr($string, $i, 2);
$i += 1;
}
} elseif ("\xE0" <= $string[$i] && $string[$i] <= "\xEF") {
$left = "\xE0" === $string[$i] ? "\xA0" : "\x80";
$right = "\xED" === $string[$i] ? "\x9F" : "\xBF";
if (!isset($string[$i+1])) {
$ret[] = $substitute;
return $ret;
} elseif ($string[$i+1] < $left || $right < $string[$i+1]) {
$ret[] = $substitute;
} else {
if (!isset($string[$i+2])) {
$ret[] = $substitute;
return $ret;
} elseif ($string[$i+2] < "\x80" || "\xBF" < $string[$i+2]) {
$ret[] = $substitute;
$i += 1;
} else {
$ret[] = substr($string, $i, 3);
$i += 2;
}
}
} elseif ("\xF0" <= $string[$i] && $string[$i] <= "\xF4") {
$left = "\xF0" === $string[$i] ? "\x90" : "\x80";
$right = "\xF4" === $string[$i] ? "\x8F" : "\xBF";
if (!isset($string[$i+1])) {
$ret[] = $substitute;
return $ret;
} elseif ($string[$i+1] < $left || $right < $string[$i+1]) {
$ret[] = $substitute;
} else {
if (!isset($string[$i+2])) {
$ret[] = $substitute;
return $ret;
} elseif ($string[$i+2] < "\x80" || "\xBF" < $string[$i+2]) {
$ret[] = $substitute;
$i += 1;
} else {
if (!isset($string[$i+3])) {
$ret[] = $substitute;
return $ret;
} elseif ($string[$i+3] < "\x80" || "\xBF" < $string[$i+3]) {
$ret[] = $substitute;
$i += 2;
} else {
$ret[] = substr($string, $i, 4);
$i += 3;
}
}
}
} else {
$ret[] = $substitute;
}
}
return $ret;
}
The result of benchmark between these functions is here.
grapheme
0.12967610359192
IntlBreakIterator::createCharacterInstance
0.17032408714294
IntlBreakIterator::createCodePointInstance
0.079245090484619
mbstring
0.081080913543701
preg_split
0.043133974075317
htmlspecialchars
0.25599694252014
byte maniplulation
0.13132810592651
The benchmark code is here.
$string = '主楼怎么走';
foreach (timer([
'grapheme' => 'str_to_array',
'IntlBreakIterator::createCharacterInstance' => 'str_to_array2',
'IntlBreakIterator::createCodePointInstance' => 'str_to_array3',
'mbstring' => 'str_to_array4',
'preg_split' => 'str_to_array5',
'htmlspecialchars' => 'str_to_array6',
'byte maniplulation' => 'str_to_array7'
],
[$string]) as $desc => $time) {
echo $desc, PHP_EOL,
$time, PHP_EOL;
}
function timer(array $callables, array $arguments, $repeat = 10000) {
$ret = [];
$save = $repeat;
foreach ($callables as $key => $callable) {
$start = microtime(true);
do {
array_map($callable, $arguments);
} while($repeat -= 1);
$stop = microtime(true);
$ret[$key] = $stop - $start;
$repeat = $save;
}
return $ret;
}

Answer 4

The documentation of str_split() says
Note:
str_split() will split into bytes, rather than characters when dealing with a multi-byte encoded string.
PHP 7.4.0 adds a new function mb_str_split(). Use this instead.

Answer 5

Assuming you have set the desired encoding and regular expression encoding for the MB functions (such as to UTF-8), you could use my method from my String class library.
/**
* Splits a string into pieces (on whitespace by default).
^
* #param string $pattern
* #param string $target
* #param int $limit
* #return array
*/
public function split(string $target, string $pattern = '\s+', int $limit = -1): array
{
return mb_split($pattern, $target, $limit);
}
By wrapping the mb_split() function in a method, I make it much easier to use. Simply invoke it with the desired value for the variable $pattern.
Remember, set the character encoding appropriately for your task.
mb_internal_encoding('UTF-8'); // For example.
mb_regex_encoding('UTF-8'); // For example.
In the case of my wrapper method, supply the empty string to the method like so.
$string = new String('UTF-8', 'UTF-8'); // Sets the internal and regex encodings.
$string->split($yourString, "")
In the direct PHP case ...
$characters = mb_split("", $string);