Match numeric substring which is preceded or followed by one of two specific substrings-CodePudding

I have a program that selects an amount from the chain that has Kč or CZK behind it. How do I edit an expression ( pattern) to check if Kč or CZK is in front of a number? See string1 and string2:

$string='Rohlík 4,99 Kč 51235';
//$string1='Rohlík CZK 4,99 51235';
//$string2='Rohlík Kč4,99 51235';

$replace = [' ', '.'];

$string = str_replace($replace,"",$string);

$string = str_replace(',',".",$string);


/*Change?*/

$pattern = '/[0-9]*[.]?[0-9]*[Kč,CZK]/';
preg_match($pattern, $string, $matches); // => 4.99 Kč
$string = $matches;

$pattern = '/[0-9]*[.]?[0-9]*/';
preg_match($pattern, $string[0], $matches);

$price = $matches[0];
print_r($price); // => 4.99

CodePudding user response：

Use logical grouping in your pattern to match the label which may come before or after the targeted number (replacing the comma with a dot can be done after this step).

Code: (Demo)

$strings = [
    'Rohlík 4,99 Kč 51235',
    'Rohlík CZK 4,99 51235',
    'Rohlík Kč4,99 51235',
    'Rohlík foo4,99 51235'
];

foreach ($strings as $string) {
    var_export(
        preg_match('/\b(?:(?:Kč|CZK) ?\K\d (?:,\d )?|\d (?:,\d )?(?= ?(?:Kč|CZK)))\b/u', $string, $m)
        ? $m[0]
        : 'not found'
    );
    echo "\n";
}

Output:

'4,99'
'4,99'
'4,99'
'not found'

Pattern Breakdown:

/                     #starting pattern delimiter
  \b                  #word boundary to guarantee matching the whole label
  (?:                 #start non-capturing group 1
    (?:Kč|CZK) ?      #non-capturing group 2 requiring one of two labels, optionally followed by a space
    \K                #forget all previously matched characters
    \d (?:,\d )?      #match the targeted integer/float value with comma as decimal placeholder
    |                 #OR
    \d (?:,\d )?      #match the targeted integer/float value with comma as decimal placeholder
    (?= ?(?:Kč|CZK))  #lookahead to for optional space followed by one of the two labels
  )                   #close non-capturing group 1
  \b                  #word boundary to guarantee matching the whole label
/                     #ending pattern delimiter
u                     #unicode/multi-byte flag