JavaScript RegExp：匹配指定分隔符前的所有符合格式条目

Let's start by figuring out why your initial regex didn't work, then fix it up properly.

Why Your First Attempt Failed #

Your regex ^[\S]*:.*(?=[\S]*\sDIVIDER) uses a positive lookahead that expects the matched line to be immediately followed by non-whitespace characters, a space, and "DIVIDER". But in your input, the *** DIVIDER sits on its own separate line after the valid actions. The lookahead can't "see" across the newline in this context, so it never triggers a match—hence why it didn't work.

Working Regex Options #

Option 1: Directly Match Pre-Divider Lines (PCRE Compatible) #

If you're using a regex engine that supports multi-line lookaheads (like PCRE), this regex will directly target only the action lines before the divider:

^[\S]+:.*$(?=(?:\n(?!\*\*\* DIVIDER).)*\n\*\*\* DIVIDER)

How it works:

• ^[\S]+:.*$: Stays true to your original pattern, matching lines that start with non-whitespace characters, followed by a colon and any content to the end of the line.
• (?=(?:\n(?!\*\*\* DIVIDER).)*\n\*\*\* DIVIDER): The positive lookahead checks that after the current line, there are zero or more lines that don't start with *** DIVIDER, followed by the actual divider line itself. This ensures we only match lines that live before the divider.

Option 2: Capture the Pre-Divider Block First (Easier to Maintain) #

If you want something more straightforward, first capture all content before the divider, then apply your original regex to that block. Use this regex to grab the pre-divider content:

\A(.*?)(?=\n\*\*\* DIVIDER)

Then run your original ^[\S]*:.*/gm regex on the captured group to pull out the action lines. This approach is often simpler to read and tweak later.

Option 3: Negative Lookbehind (For Modern Engines) #

If your regex engine supports variable-length lookbehinds (like JavaScript ES2018+ or PCRE), this regex will exclude any lines that come after the divider:

^[\S]+:.*$(?<!\*\*\* DIVIDER\n.*)

It works by checking that the current line isn't preceded by the divider line and any content that comes after it.

Testing with Your Input #

For your sample input:

john: buys beer
sam: eats bananas
*** DIVIDER
irrelevant actions
luis: walking

All valid options will only match these two lines:

• john: buys beer
• sam: eats bananas

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