strcoll()'s behavior is sometimes a little bit confusing. It depends on LC_COLLATE in your locale.
<?php
$a = 'a';
$b = 'A';
print strcmp ($a, $b) . "\n"; // prints 1
setlocale (LC_COLLATE, 'C');
print "C: " . strcoll ($a, $b) . "\n"; // prints 1
setlocale (LC_COLLATE, 'de_DE');
print "de_DE: " . strcoll ($a, $b) . "\n"; // prints -2
setlocale (LC_COLLATE, 'de_CH');
print "de_CH: " . strcoll ($a, $b) . "\n"; // prints -2
setlocale (LC_COLLATE, 'en_US');
print "en_US: " . strcoll ($a, $b) . "\n"; // prints -2
?>
This is useful e. g. if want to sort an array by using strcoll:
<?php
$a = array ('a', 'A', 'ä', 'Ä', 'b', 'B');
setlocale (LC_COLLATE, 'C');
usort ($a, 'strcoll');
print_r ($a);
?>
This is like sort($a):
Array
(
[0] => A
[1] => B
[2] => a
[3] => b
[4] => Ä
[5] => ä
)
<?php
setlocale (LC_COLLATE, 'de_DE');
usort ($a, 'strcoll');
print_r ($a)
?>
This is completely different:
Array
(
[0] => a
[1] => A
[2] => ä
[3] => Ä
[4] => b
[5] => B
)
strcoll
Description
int strcoll ( string str1, string str2)Returns < 0 if str1 is less than str2; > 0 if str1 is greater than str2, and 0 if they are equal. strcoll() uses the current locale for doing the comparisons. If the current locale is C or POSIX, this function is equivalent to strcmp().
Note that this comparison is case sensitive, and unlike strcmp() this function is not binary safe.
Not: strcoll() was added in PHP 4.0.5, but was not enabled for win32 until 4.2.3.
See also ereg(), strcmp(), strcasecmp(), substr(), stristr(), strncasecmp(), strncmp(), strstr(), and setlocale().
strcoll
sakkarinlaohawisut15 at hotmail dot com
22-Mar-2003 11:31
22-Mar-2003 11:31
27-Aug-2002 07:05
Note that some platforms implement strcmp() and strcasecmp() according to the current locale when strings are not binary equal, so that strcmp() and strcoll() will return the same value! This depends on how the PHP strcmp() function is compiled (i.e. if it uses the platform specific strcmp() found in its standard library!).
In that case, the only difference between strcoll() and strcmp() is that strcoll() may return 0 for distinct strings(i.e. consider strings are equal) while strcmp() will differentiate them if they have distinct binary encoding! This typically occurs on Asian systems.
What you can be sure is that strcmp() will always differentiate strings that are encoded differently, but the relative order may still use the current locale setting for collation order!
