preg_replace在模式之间删除字符串的多个不同部分

Question

This is a very small portion of a string i have scraped:

SEPVALAjazz(music)SEPVALC5.59SEPVALAlatin(music)SEPVALAcountry(music)SEPVALC6.70SEPVALArock(music)SEPVALC5.89SEPVALAdance(music)SEPVALAworld(music)SEPVALC6.70

I have manipulated the string with regex to insert the "SEPVALA" and "SEPVALC" values. I wish to remove the string between the patterns "SEPVALA" and "(music)SEPVALA" where ever it occurs in the text. In the above example this would remove "latin" and "dance".

The desired result would be:

SEPVALAjazz(music)SEPVALC5.59SEPVALAcountry(music)SEPVALC6.70SEPVALArock(music)SEPVALC5.89SEPVALAworld(music)SEPVALC6.70

The code i am implementing at the moment is:

$pattern="/SEPVALA(.+?)\(music\)SEPVALA/";

$replacement="SEPVALA";

$newstring=preg_replace($pattern, $replacement, $originalstring);

However this removes too much of the string as the regex engine picks up the very first "SEPVALA" and removes everything forward from this point until "(music)SEPVALA" (as it should do!) I need a method to scan backwards from "(music)SEPVALA" removing the offending text and stopping at the next "SEPVALA" before then moving forwards to the next match of "(music)SEPVALA" and so on. Or any other method to accomplish this. Many thanks.

Answer 1

Try replacing the following pattern with an empty string:

SEPVALA((?!SEPVAL[AC]).)+(?<=\(music\))(?=SEPVALA)

Example: http://www.rubular.com/r/ZCfJQ4ey2O

Explanation: This works by matching SEPVALA, then matching one character at a time until either SEPVALA or SEPVALC can match, by using a negative lookahead with repetition. At this point we have matched from SEPVALA to just before either SEPVALA or SEPVALC, so we just need to make sure the previously matched characters are (music), and the next characters are SEPVALA.

Full code:

$pattern="/SEPVALA((?!SEPVAL[AC]).)+(?<=\\(music\\))(?=SEPVALA)/";
$replacement="";
$newstring=preg_replace($pattern, $replacement, $originalstring);