正则表达式重复捕获组并忽略指定字符串问题求助

Let's break down what's wrong with your current regex and fix it to capture both Stream #0:4 and Stream #0:5 while ignoring the intermediate -> #0:0 (copy) content.

What's Wrong with the Original Regex? #

Your regex /(?<=Stream mapping:)(\s+)([a-zA-Z])+(\s+)#([0-9]:[0-9])/s has a few key issues:

• The ([a-zA-Z])+ capture group only captures the last letter of the word it matches (in this case, "m" from "Stream") because the + quantifier makes the group repeat, but only stores the final iteration.
• It doesn't account for the intermediate -> #x:x (copy) content that sits between your target entries.
• It captures unnecessary whitespace groups, cluttering your results.

The Solution #

To capture all Stream #x:x entries after Stream mapping: and ignore the copy-related content, use a global match on the pattern that directly targets your desired format.

Regex Pattern

/Stream #\d:\d/g

How It Works

• Stream #: Matches the literal prefix of your target entries.
• \d:\d: Matches the digit-colon-digit format (e.g., 0:4, 0:5).
• The g modifier enables global matching, so it finds all occurrences of the pattern in your text.

Example Usage (PHP)

$input = "Stream mapping: Stream #0:4 -> #0:0 (copy) -> #0:1 (copy) Stream #0:5";
preg_match_all('/Stream #\d:\d/', $input, $matches);
print_r($matches[0]);

Output

Array
(
    [0] => Stream #0:4
    [1] => Stream #0:5
)

If You Need a Single Regex with Capture Groups (For Non-Global Scenarios)

If you want a regex that directly captures both entries in groups (without relying on global matching), use this pattern with the s modifier (to make . match newlines):

/(?<=Stream mapping:).*?(Stream #\d:\d).*?(Stream #\d:\d)/s

• (?<=Stream mapping:): Ensures we start matching right after the "Stream mapping:" prefix.
• .*?: Non-greedily matches any characters (including newlines, thanks to s) until it hits the first target entry.
• (Stream #\d:\d): Captures the first target entry.
• The second .*? skips the intermediate copy content, and the final capture group grabs the second entry.

This will give you exactly the two captured groups you're looking for.

内容的提问来源于stack exchange，提问作者Bruno Andrade