Question

Suppose I have a string, like so:

$string = 'president barack obama';

Now, suppose I want to explode this into an array, breaking at the words. You’d think I can just use explode(), right? That works, but what if I want an array of all possible left-to-right combinations of the words? Like so:

Array
(
    [0] => 'barack'
    [1] => 'barack obama'
    [2] => 'obama'
    [3] => 'president'
    [4] => 'president barack'
    [5] => 'president barack obama'
)

What is the most efficient way to do this?

---

Possible solution:

I’ve come up with one possible solution so far, but I’m hoping one of you can give me a better idea. I imagine approaching this like so:

1. Explode normally.
2. Loop through each word.
3. For each word, store it in an array. Then, check if there is another word in the array (after itself). If there is, add a new array value which consists of $current_word . ' ' . $new_word;. Do this for each word.

Now, that will probably work. However, it seems annoying, and I’m afraid someone else may have a better way of doing this. What do you all recommend? Is there, perhaps, a PHP function that does just this that I don’t know about?

Answer 1

This should do what you are looking for:

$string  = 'president barack obama won';
$results = getWordPermutaions($string);
print_r($results);

function getWordPermutaions($inStr) {
  $outArr   = Array();
  $tokenArr = explode(" ", $inStr);
  $pointer  = 0;
  for ($i=0; $i<count($tokenArr); $i++) {
    $outArr[$pointer] = $tokenArr[$i];
    $tokenString = $tokenArr[$i];
    $pointer++; 
    for ($j=$i+1; $j<count($tokenArr); $j++) {
      $tokenString .= " " . $tokenArr[$j];
      $outArr[$pointer] = $tokenString;
      $pointer++;
    }
  }
  return $outArr;
}

/*
$results:
Array (
        [0] => 'president'
        [1] => 'president barack'
        [2] => 'president barack obama'
        [3] => 'president barack obama won'
        [4] => 'barack'
        [5] => 'barack obama'
        [6] => 'barack obama won'
        [7] => 'obama'
        [8] => 'obama won'
        [9] => 'won'
)
*/