PHP substring matching whole words

Question

I am trying write a StringMatch function that returns true when words from one string can be found in another string. The exception is that I don't want matches for things like plurals and other suffixes, and I would also like to avoid matching when a word is prefixed.

To explain more visually:

apple watch - apple watches (no match)
apple watch - apple watch repairs (match)
apple watch - new apple watch (match)
apple watch - pineapple watch (no match)

I would like is this:

echo StringMatch("apple watch", "apple watches");       // output 0
echo StringMatch("apple watch", "apple watch repairs"); // output 1
echo StringMatch("apple watch", "new apple watch");     // output 1
echo StringMatch("apple watch", "pineapple watch");     // output 0

I have had some basic success with strpos() I cannot figure out how to return "0" when the second string contains suffixes or prefixes as per examples above.

Here is how I'm trying to solve it:

function StringMatch($str1,$str2)
{
    if (SomeFunctionOrRegex($str1,$str2) !== false)
    {
        return(1);
    }
    else
    {
        return(0);
    }
}

Perhaps there is a graceful regex solution. I have tried strpos() but it is not strict enough for my needs.

Answer 1

Like this as I said in the comments

function StringMatch($str1,$str2)
{
  return preg_match('/\b'.preg_quote($str1,'/').'\b/i', $str2);
}

echo StringMatch("apple watch", "apple watches");       // output 0
echo "\n";
echo StringMatch("apple watch", "apple watch repairs"); // output 1
echo "\n";
echo StringMatch("apple watch", "new apple watch");     // output 1
echo "\n";
echo StringMatch("apple watch", "pineapple watch");     // output 0
echo "\n";

Output:

0
1
1
0

Sandbox

Preg Quote in necessary to avoid issues where $str1 could contain things like . which in Regex is any character.

Furthermore you could strip punctuation like this

$str1 = preg_replace('/[^\w\s]+/', '', $str1);

For example:

echo StringMatch("apple watch.", "apple watch repairs"); // output 1

Without removing the punctuation, this will return 0. Rather or not that is important is up to you.

Sandbox

UPDATE

Match out of order, for example:

//words out of order
echo StringMatch("watch apple", "new apple watch");     // output 1

The easy way is implode/explode:

function StringMatch($str1,$str2)
{
  //use one or the other
  $str1 = preg_replace('/[^\w\s]+/', '', $str1);
  //$str1 = preg_quote($str1,'/');
  $words = explode(' ', $str1);
  preg_match_all('/\b('.implode('|',$words).')\b/i', $str2,     $matches);
  return count($words) == count($matches[0]) ? '1' : '0';
}

Sandbox

You can also skip the explode/implode and use

 $str1 = preg_replace('/\s/', '|', $str1);

Which can be combined to the other preg_replace

 $str1 = preg_replace(['/[^\w\s]+/','/\s/'], ['','|'], $str1);

Or all together

function StringMatch($str1,$str2)
{
  $str1 = preg_replace(['/[^\w\s]+/','/\s/'], ['','|'], $str1);
  preg_match_all('/\b('.$str1.')\b/i', $str2, $matches);
  return (substr_count($str1, '|')+1) == count($matches[0]) ? '1' : '0';
}

Sandbox

But then of course you can't count the words array, but you can count the number of | pipes which is 1 less then the number of words (hence the +1). That is if you care that all the words match.