Regex to match if given text is not found and match as little as possible

问题

I have text:

<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>
<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>
<a>
sdfsdf
<b>BBzz</b>
sdfsdf
</a>
<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>

I can't parse it as xml. I need to use regex here. Also this is only example.

I want regex that can match every group <a>...</a> that does not contain element b with text that starts with BB.

I came up with this regex: <a>.*?<b>(?!B).*?</b>.*?</a> But it matches last group as:

<a>
sdfsdf
<b>BBzz</b>
sdfsdf
</a>
<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>

Which is bad for me.

How to write regex that will only match those 3 group from my given example?:

1.

<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>

2.

<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>

3.

<a>
sdfsdf
<b>DDzz</b>
sdfsdf
</a>

回答1:

Use a tempered greedy token regex:

<a>(?:(?!<(?:b>BB|/?a>)).)*</a>

Enable the . matches newline option.

Details:

• <a> - a literal <a> char sequence
• (?:(?!<(?:b>BB|/?a>)).)* - a tempered greedy token matching any char (.) that is not the starting symbol of a sequence that can be matched with the pattern inside the (?!<(?:b>BB|/?a>)) lookahead (not a <b>BB or </a> or <a> sequence)
• </a> - a literal </a> char sequence

来源：https://stackoverflow.com/questions/40860847/regex-to-match-if-given-text-is-not-found-and-match-as-little-as-possible