str_replace
I was reminded again of the desire for a generic str_rot function. Character manipulation loops in PHP are slow compared to their C counterparts, so here's a tuned version of the previous function I posted. It's 1.6 times as fast, mainly by avoiding chr() calls.
<?php
function str_rot($s, $n = 13) {
static $letters = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ';
$n = (int)$n % 26;
if (!$n) return $s;
if ($n == 13) return str_rot13($s);
for ($i = 0, $l = strlen($s); $i < $l; $i++) {
$c = $s[$i];
if ($c >= 'a' && $c <= 'z') {
$s[$i] = $letters[(ord($c) - 71 + $n) % 26];
} else if ($c >= 'A' && $c <= 'Z') {
$s[$i] = $letters[(ord($c) - 39 + $n) % 26 + 26];
}
}
return $s;
}
?>
But using strtr() you can get something 10 times as fast as the above :
<?php
function str_rot($s, $n = 13) {
static $letters = 'AaBbCcDdEeFfGgHhIiJjKkLlMmNnOoPpQqRrSsTtUuVvWwXxYyZz';
$n = (int)$n % 26;
if (!$n) return $s;
if ($n < 0) $n += 26;
if ($n == 13) return str_rot13($s);
$rep = substr($letters, $n * 2) . substr($letters, 0, $n * 2);
return strtr($s, $letters, $rep);
}
?>
This technique is faster because PHP's strtr is implemented in C using a byte lookup table (it has O(m + n) complexity). However, PHP 6 will use Unicode, so I guess(?) strtr will then have to be implemented with a search for each character (O(m * n)). Using strtr might still be faster since it offloads the character manipulation to C rather than PHP, but I don't really know. Take your pick.
Happy coding!
(Benchmark code):
<?php
for ($k = 0; $k < 10; $k++) {
$s = 'The quick brown fox jumps over the lazy dog.';
$t = microtime(1);
for ($i = 0; $i < 1000; $i++) $s = str_rot($s, $i);
$t = microtime(1) - $t;
echo number_format($t, 3) . "\n";
}
?>
str_rot13
(PHP 4 >= 4.2.0, PHP 5)
str_rot13 — Perform the rot13 transform on a string
Description
string str_rot13
( string
$str
)
Performs the ROT13 encoding on the str argument and
returns the resulting string.
The ROT13 encoding simply shifts every letter by 13 places in the alphabet while leaving non-alpha characters untouched. Encoding and decoding are done by the same function, passing an encoded string as argument will return the original version.
Parameters
-
str -
The input string.
Return Values
Returns the ROT13 version of the given string.
Examples
Example #1 str_rot13() example
<?php
echo str_rot13('PHP 4.3.0'); // CUC 4.3.0
?>
Changelog
| Version | Description |
|---|---|
| 4.3.0 |
The behaviour of this function was fixed. Before this fix, the
str was also modified, as if it was passed
by reference.
|
add a note
User Contributed Notes
str_rot13 - [9 notes]
shaun ¶
1 year ago
peter at NOSPAM jamit dot com ¶
3 years ago
This ROT13 variant is different from my earlier version in that it retains 'ethnicity'. For example, a Chinese text when encrypted will remain Chinese, and the string will not be making sense (the real meaning will be encrypted). Just look at the code and you will understand.
<?php
function unichar2ords($char, $encoding = 'UTF-8') {
$char = mb_convert_encoding($char, 'UCS-4', $encoding);
$val = unpack('N', $char);
return $val[1];
}
function ords2unichar($ords, $encoding = 'UTF-8'){
$char = pack('N', $ords);
return mb_convert_encoding($char, $encoding, 'UCS-4');
}
function mbStringToArray ($string, $encoding = 'UTF-8') {
if (empty($string)) return false;
for ($strlen = mb_strlen($string, $encoding); $strlen > 0; ) {
$array[] = mb_substr($string, 0, 1, $encoding);
$string = mb_substr($string, 1, $strlen, $encoding);
$strlen = $strlen - 1;
}
return $array;
}
function unicodeRotN($str, $offset, $encoding = 'UTF-8') {
$val = '';
$array = mbStringToArray ($str, $encoding = 'UTF-8');
$len = count($array);
for ($i = 0; $i < $len; $i++) {
$val .= ords2unichar(unichar2ords($array[$i], $encoding) + $offset, $encoding);
}
return $val;
}
// example
$original = '中國是我的家'; // means "China is my home"
$encrypted = unicodeRotN($string, 13); // 为團昼戞皑寃 means "Ñ Ai injustice for the Mission Day" (Google translation)
$decrypted = unicodeRotN($encrypted, -13); // 中國是我的家
?>
peter at NOSPAM jamit dot com ¶
3 years ago
Here is my ROT13 function that works for all possible characters and not just ASCII. It can be used on Chinese, Japanese, ....
<?php
function rot13encrypt ($str) {
return str_rot13(base64_encode($str));
}
function rot13decrypt ($str) {
return base64_decode(str_rot13($str));
}
// example
$string = '中國是我的家';
$encrypted = rot13encrypt ($string); // produces 5Yvg5MlY5cvi5bvE55dR5n62
$decrypted = rot13decrypt ($encrypted); // produces 中國是我的家
?>
electro at whatever dot com ¶
5 years ago
<?php
/**
* Rotate each string characters by n positions in ASCII table
* To encode use positive n, to decode - negative.
* With n = 13 (ROT13), encode and decode n can be positive.
*
* @param string $string
* @param integer $n
* @return string
*/
function rotate($string, $n) {
$length = strlen($string);
$result = '';
for($i = 0; $i < $length; $i++) {
$ascii = ord($string{$i});
$rotated = $ascii;
if ($ascii > 64 && $ascii < 91) {
$rotated += $n;
$rotated > 90 && $rotated += -90 + 64;
$rotated < 65 && $rotated += -64 + 90;
} elseif ($ascii > 96 && $ascii < 123) {
$rotated += $n;
$rotated > 122 && $rotated += -122 + 96;
$rotated < 97 && $rotated += -96 + 122;
}
$result .= chr($rotated);
}
return $result;
}
$enc = rotate('string', 6);
echo "Encoded: $enc<br/>\n";
echo 'Decoded: ' . rotate($enc, -6);
?>
maximius at gmail dot com ¶
5 years ago
Perhaps someone will find this useful ;)
<?
function rotN($s, $n){
$s2 = "";
for($i = 0; $i < strlen($s); $i++){
$char2 = $char = ord($s{$i});
$cap = $char & 32;
$char &= ~ $cap;
$char = $char > 64 && $char < 123 ? (($char - 65 + $n) % 26 + 65) : $char;
$char |= $cap;
if($char < 65 && $char2 > 64 || ($char > 90 && $char < 97 && ($char2 < 91 || $char2 > 96))) $char += 26;
else if($char > 122 && $char2 < 123) $char -= 52;
if(strtoupper(chr($char2)) === chr($char2)) $char = strtoupper(chr($char)); else $char = strtolower(chr($char));
$s2 .= $char;
}
return $s2;
}
?>
It takes any string, $s, and any ROT value, $n. Just like str_rot13, it's both an encoder and decoder. To decode an encoded string, just pass -$n instead of $n.
shaunspiller at spammenotgmail dot com ¶
3 years ago
Here's my implementation of a str_rot that takes a custom offset. It's faster than the others here because it allocates the output string in one go instead of tacking on characters one at a time. It can handle positive or negative offsets of any size, and it fixes everything up to ensure only upper and lower case letters are translated and that they wrap around correctly within the alphabet.
<?php
function str_rot($s, $n = 13) {
$n = (int)$n % 26;
if (!$n) return $s;
for ($i = 0, $l = strlen($s); $i < $l; $i++) {
$c = ord($s[$i]);
if ($c >= 97 && $c <= 122) {
$s[$i] = chr(($c - 71 + $n) % 26 + 97);
} else if ($c >= 65 && $c <= 90) {
$s[$i] = chr(($c - 39 + $n) % 26 + 65);
}
}
return $s;
}
?>
grawity+phpnet at gmail dot com ¶
5 years ago
resubmitting updated function in re:#76975
<?php
function asc_shift($str,$offset=0) {
$new = '';
for ($i = 0; $i < strlen($str); $i++)
$new .= chr(ord($str[$i])+$offset);
return $new;
}
?>
because as of PHP 6, $str{$i} is deprecated.
arwab at surrealwebs dot com ¶
5 years ago
here's my rot function, it works anyway
<?php
/**
* preforms the rotation algorithm on the passed in string
*/
function _rot( $str , $dist=13 ){
if( !is_numeric($dist) || $dist < 0){
$dist = 13;
}
$u_lower = 65; $u_upper = 90;
$l_lower = 97; $l_upper = 122;
$char_count = ($u_upper - $u_lower) +1;
while( $dist > $char_count ){
$dist -= $char_count;
}
$newstr = '';
for( $i=0; $i<strlen($str); ++$i){
$c = ord($str[$i]);
/*
* Check if the character is within the bounds of our function (a-zA-z)
* if not it gets tacked on to the string as is and we move on to the
* next one.
*/
if( $c<$u_lower || $c>$l_upper || ( $c>$u_upper && $c <$l_lower ) ){
$newstr .= chr($c);
continue;
}
$lower = ( $c<=$u_upper?$u_lower:$l_lower);
$upper = ( $c<=$u_upper?$u_upper:$l_upper);
$c += $dist;
if( $c > $upper){
$c = (($c - $upper) + ($lower-1));
}
$newstr .= chr($c);
}
return $newstr;
}
?>
nick at lazy-river dot net ¶
5 years ago
This is recursive function to shift the component letters of a string left or right in the ascii table.
I've left it simple as it suits my needs, but you may want to include error checking for a null string and also put bounds in place, or make it actually rotate around the whole character set rather than just shifting the string up or down.
function asc_shift($string, $amount) {
$key = substr($string, 0, 1);
if(strlen($string)==1) {
return chr(ord($key) + $amount);
} else {
return chr(ord($key) + $amount) . asc_shift(substr($string, 1, strlen($string)-1), $amount);
}
}
For example:
<?php
echo asc_shift("TESTING12345@", 5);
?>
shifts every character up 5 ascii positions, resulting in this string:
YJXYNSL6789:E
In reverse:
<?php
echo asc_shift("YJXYNSL6789:E", -5);
?>
shifts every character down 5 ascii positions, resulting in this string:
TESTING12345@