php str_replace replacing itself

Question

I need to replace every occurrence of one of the letters a,o,i,e,u with [aoieu]?
I tried to do the following:

str_replace(array('a', 'o', 'i', 'e', 'u'), '[aoieu]?', $input);

But when giving it input of black instead of giving me the expected bl[aoieu]?ck it gave me

bl[a[ao[aoi[aoie[aoieu]?]?[aoieu]?]?[aoie[aoieu]?]?[aoieu]?]?[aoi[aoie[aoieu]?]?[aoieu]?]?[aoie[aoieu]?]?[aoieu]?]?ck

How can I get it to not replace things it already replaced?

YoYoMa's comment: "Apparently some have missed this warning @ http://us3.php.net/manual/en/function.str-replace.php Caution Replacement order gotcha Because str_replace() replaces left to right, it might replace a previously inserted value when doing multiple replacements. See also the examples in this document. Others have answered the question with possible solutions to the problem. This is why the problem is occurring." — Peter O., Feb 14 '12 at 20:17

Emil Vikström · Accepted Answer · 2011-08-11T13:38:13.853

5

You can consider using a regular expression for this, or you can make your own function which steps through the string one letter at a time. Here's a regex solution:

preg_replace('/[aoieu]/', '[aoieu]?', $input);

Or your own function (note that $search only can be a single char or an array of chars, not strings - you can use strpos or similar to build one which handles longer strings as well):

function safe_replace($search, $replace, $subject) {
  if(!is_array($search)) {
    $search = array($search);
  }
  $result = '';
  $len = strlen($subject);
  for($i = 0; $i < $len; $i++) {
    $c = $subject[$i];
    if(in_array($c, $search)) {
      $c = $replace;
    }
    $result .= $c;
  }
  return $result;
}
//Used like this:
safe_replace(array('a', 'o', 'i', 'e', 'u'), '[aoieu]?', 'black');

edited Aug 11 '11 at 13:38

answered Aug 11 '11 at 13:32

Emil Vikström

90,431
16
141
175

The regexp one is working. about the second one, I'm not a fan of 'do it your own' after all the problems it can do with coding, especially with interpreted language. – Daniel Aug 11 '11 at 13:41
What problems are you talking about? Performance problems, debugging problems, 'bad programmer == bad code' problems or something else? – Emil Vikström Aug 11 '11 at 13:46
1

the kind of problems when you implement something thats already implemented - a bunch of useless code. also there are performance issues, string manipulation is slow in non native code. – Daniel Aug 11 '11 at 14:12
same problem here but. trying to use the function but not sure what i should be adding for the 2nd variable currently '[aoieu]?'? – v3nt Sep 18 '12 at 12:30

score 3 · Answer 2 · answered Aug 11 '11 at 13:30

3

You might want to try this

<?php
$string = 'black';
$pattern = '/([aeiou])/i';
$replacement = '[aeiou]';
echo preg_replace($pattern, $replacement, $string);
?>

answered Aug 11 '11 at 13:30

az4dan

651
2
10
30

score 3 · Answer 3 · answered Aug 11 '11 at 13:31

3

Taken from the documentation:

Replacement order gotcha

Because str_replace() replaces left to right, it might replace a previously inserted value when doing multiple replacements.

answered Aug 11 '11 at 13:31

powtac

40,542
28
115
170

score 3 · Answer 4 · edited Apr 12 '21 at 23:54

I recommend avoiding preglike functions and using strtr().

This native function

makes a single pass over the input string,
does not replace replacements, and
finds the longest matching substring to replace (when a qualifying string is found within another qualifying string)

Code:

$result = strtr($input, array('a' => '[aoieu]?', 
                         'o' => '[aoieu]?', 
                         'i' => '[aoieu]?', 
                         'e' => '[aoieu]?', 
                         'u' => '[aoieu]?'));

score 0 · Answer 5 · answered Aug 11 '11 at 13:32

0

$input = str_replace(array('a', 'o', 'i', 'e', 'u'),   '~',          $input);
$input = str_replace('~',                              '[aoieu]?',   $input);

answered Aug 11 '11 at 13:32

powtac

40,542
28
115
170

1

Fails when you have a `~` in the input. Trivial reason that this is not a robust solution. – Lightness Races in Orbit Aug 11 '11 at 13:33
You cannot "find another pattern which does not occur" in the general case. And you should _prove_ that it'll be fastest if you want to assert that! – Lightness Races in Orbit Aug 11 '11 at 13:37
It won't be fastest probably, but using a weird character like `¯` or `œ` might do the trick. – Madara's Ghost Aug 11 '11 at 13:40
http://www.google.com/search?q=str_replace+vs.+preg_replace lot of benchmarks for str_replace vs. preg_replace out there. – powtac Aug 11 '11 at 13:48

score 0 · Answer 6 · answered Aug 11 '11 at 13:34

0

Here it is:

$output = preg_replace('/[aeiou]/', '[aeiou]?', $input);

answered Aug 11 '11 at 13:34

Samnan

653
1
5
13

score 0 · Answer 7 · answered Aug 11 '11 at 13:34

You might be able to get preg_replace to handle this for you (see Thax, Emil, etc.'s answers). Otherwise, if that is too complicated, you can, tokenize:

$token = '{{{}}}';
// replace the temporary value with the final value
str_replace( $token, '[aoieu]?', 
    // replace all occurances of the string with a temporary value.
    str_replace( (array('a', 'o', 'i', 'e', 'u'), $token, $input ) );

php str_replace replacing itself

7 Answers7

Linked