用PHP多行文本正则表达式头痛

Question

I've unformatted text data extracted from pdf like this:

AB01234 This could be a

long question with multiple

new lines a)these b)are c)the responses which could

contains new lines d)either b

AB01235 This is another question with same multiple

response a) one b) two c) three d) four c

...

My goal is to group questions identifiers, questions, answers and correct answer which is the last char. Is there any way to do so with a regex?

{
   [0] => 'AB01234',
   [1] => 'This could be a long question with multiple new lines',
   [2] => 'these'
   [3] => 'are',
   [4] => 'the responses which could contains new lines',
   [5] => 'either',
   [6] => 'b'
}

Answer 1

I would not attempt to do this with a single regular expression. There is far too much variance in the input. I would clean up the text like this:

$text = '
    AB01234 This could be a
    long question with multiple
    new lines a)these b)are c)the responses which could
    contains new lines d)either b
    AB01235 This is another question with same multiple
    response a) one b) two c) three d) four c
';
$text = preg_replace('/([A-Z]{2}[0-9]{5})/', ' QUESTION\1 ', $text);
$text = preg_replace('/([a-z]\))/', ' ANSWER\1 ', $text);
$text = trim(preg_replace('/\s+/', ' ', $text));
print($text);

You will see that the text is now rather clean. It is one line. Spacing is cleaned up. You also have clear flags for QUESTION and ANSWER. You can change those to anything you like, such as !@#$#@!# for a question. They just have to be something that will not ever appear in the text.

Now, you could try a regular expression, but explode is easier at this point because you flagged the delimiters. I used explode and implode a lot in this example, just in case you haven't seen it much. You don't have to use it. You could use regex or substrings.

$questions = array();
$qas = explode("QUESTION", $text);
foreach($qas as $qa)
{
    if($qa == "") continue;
    $answers = explode("ANSWER", $qa);
    $q = array();
    foreach($answers as $i=>$answer)
    {
        $a = explode(' ', $answer);
        if($i == 0) $q[] = $a[0];
        $questions[0] = $a[0];
        array_shift($a);
        $q[] = implode(' ', $a);
    }
    $questions[] = $q;
}
print_r($questions);

Now, you should have an array that you want.