Given a string such as

cite{AA,BBB, C}  skip{DD}   cite{EE,F}

, I am trying to extract the comma-separated strings in the specific tag (in this case, named as cite) using Regex. Thus, the output for the above string should result in

AA
BBB
C
EE
F

I think /cite{(. ?)}/ selects strings in \cite{....} form, but how do we then split the string in this output? I find that some expressions like [^,(?! )] split the strings based on comma, but I cannot find a way to couple these two things.

CodePudding user response：

You could split on the comma using a lookbehind, checking for the cite{ prefix, for example:

• 

  But outside of using a programming language, I'm not sure how you would capture what you want if it's not fixed length, in a single regex. Though if it is fixed-width, you could do something like this:

  • 

    CodePudding user response：

    Suppose the string were

    cite{AA,BBB, C}  skip{DD} GG  cite{EE,F}
         ^^ ^^^  ^                     ^^ ^
    

    We want to match the five substrings indicated by the carets.

    If the regex engine supports variable-length lookbehinds we could use the regular expression

    (?<=\bcite{[^{}]*?)[A-Z] (?=[^{}]*})
    

    Demo (click the "Context" tab at the link to see the matches).

    This regex can be broken down as follows.

    (?<=         # begin a positive lookbehind 
      \bcite     # match literal preceded by a word break
      {          # match character
      [^{}]*?    # match zero or more chars other than braces, lazily
    )            # end positive lookbehind
    [A-Z]        # match >= 1 uppercase letters
    (?=          # begin positive lookahead
      [^{}]*?    # match zero or more chars other than braces, then '}'    
    )            # end positive lookahead