How to extract substrings that follow specific keywords in a string?

Question

I need to process strings that have a mostly regular format/structure. Basically, the string contains 3 keywords that always appear in the same order: ATLPáscoa, ATLNatal, and ATLVerão

Between these keywords are an unknown number of whitespace characters. Also, there is the possibility that each of the keywords will be followed by a date value that may consist of non-whitespace and whitespace characters.

Associated by their keyword, I want to declare 3 variables called $datePáscoa, $dateNatal, and $dateVerão and assign the date substring to these variables.

Here's an example:

$string = 'ATLPáscoa            ATLNatal          ATLVerão     Turno11-03a07desetembro';

My desired output is:

$datePáscoa = '';
$dateNatal = '';
$dateVerão = 'Turno11-03a07desetembro';

Here is another example:

$string = 'ATLPáscoa  bananas   ATLNatal xyza sd af          ATLVerão      Turno11-03a07desetembro';

My expected output is:

$datePáscoa = 'bananas';
$dateNatal = 'xyza sd af';
$dateVerão = 'Turno11-03a07desetembro';

I tried to use the str_replace(), but it is clearly not the way:

$string = str_replace("Atelier","",$string );
$string = str_replace("Páscoa","",$string );
$string = str_replace("Natal","",$string );
$string = str_replace("Verão","",$string );

How can I extract the date values and assign the values to the appropriate variable?

Answer 1

Code: (Demo <- with an alternative input string)

$string = 'ATLPáscoa  banana   ATLNatal xyza sd af          ATLVerão      Turno11-03a07desetembro';

$datePáscoa = preg_match('~ATLPáscoa\s*\K(?!ATL)\S+(?:\s+(?!ATL)\S+)*~u', $string, $out) ? $out[0] : '';
$dateNatal = preg_match('~ATLNatal\s*\K(?!ATL)\S+(?:\s+(?!ATL)\S+)*~u', $string, $out) ? $out[0] : '';
$dateVerão = preg_match('~ATLVerão\s*\K\S+(?:\s+\S+)*~u', $string, $out) ? $out[0] : '';

echo '$datePáscoa = '; var_export($datePáscoa); echo "\n";
echo '$dateNatal = '; var_export($dateNatal); echo "\n";
echo '$dateVerão = '; var_export($dateVerão);

Output:

$datePáscoa = 'banana'
$dateNatal = 'xyza sd af'
$dateVerão = 'Turno11-03a07desetembro'

If this were my project, I'd probably build a single regex function call that returns all of the matches in an array, then I'd extract what I wanted, when I wanted it. You have asked for individually named variables, so I think 3 function calls will be simplest to demonstrate.

The input that you have offered doesn't require the inclusion of the u pattern modifier, but I am adding it in case your actual data requires it.

\K tells the regex engine to "release previously matched characters" from the fullstring match -- this is used to avoid the use of a capture group and ensure your returned value is only the "white meat". The same reason is why you see \S+(?:\s+\S+)* -- which matches a "word" then optionally matches one or more whitespaces followed by another "word".

I am using var_export() in my demo to show that there are no leading or trailing whitespace characters in the results.

(?!ATL) in the first two patterns is used to avoid "over matching" or basically "matching too far". The third pattern doesn't require this consideration.

Answer 2

Well, since ATL is in every part, I would start with an explode:

$array = explode("ATL", $string);

Note that $array[0] will be an empty string (in this case, but as I see you get the same type of input every time), and then trim the leading and trailing spaces like this:

for ( $i = 0; $i < count($array); $i++ ) {
 trim($array[$i]);
}

And then copying into their respective variables:

$datePáscoa = $array[1];
$dateNatal = $array[2];
$dateVerão = $array[3];

At this point, they still contain their names, so we cut it with a combo of strpos() (which gives back the position of a specific string) and strstr() (returns part of the string from a given pointer), like:

$datePáscoa = strstr($datePáscoa,strpos($datePáscoa," ")+1);

That's a whitespace there. After that maybe trim them again, as they may still contain some whitespaces after the strstr.