Javascript正则表达式：相同模式中匹配和不匹配的混合

Question

I had been working with a few regular expressions and I ran into a problem.

I need to run a replace on the following pattern:

The string must have a \\E\\ but it must not be preceded by \\.br or a \\R

Sample text:

Random\E\Text  - should match

\E\Text        - should match

\R\E\.br\      - should not match

\.br\E\R\      - should not match

Could anyone suggest a pattern that would be able to accomplish this?

I tried only for the \\.br case and while

/[^\b\\\.br\b]\E\/g

fails terribly with good reason, the following pattern seemed to not work either:

/(?!\\\.br)\E\/g

Any help would be appreciated.

Thanks.

Answer 1

You could try (\\\\.br|\\\\R)\\\\E(\\\\.br|\\\\R)

Javascript doesn't support lookbehinds, but you should be able to negatively assert a match (if !match, string is valid).

Working example: http://regex101.com/r/hL3wN5

var str = "Random\\E\\Text";
if (!(/(\\.br|\\R)\\E(\\.br|\\R)/g).test(str))
    //do stuff because it doesn't test true

Answer 2

Unfortunately, JavaScript does not support lookbehinds, only lookaheads. However, all is not lost! You can a function for replacement:

input.replace( /(...|..|.|$)\\E\\/, function( match, prefix ) {
    if( prefix==='\.br' || prefix.match( /\\R$/ ) )
        return match;  // return match unaffected
    else
        return prefix + '<\\E\\ replacement>';
});

I generally find that a lot of regex problems can be solved by this kind of approach: instead of doing everything in one giant regex, break it up into manageable chunks, and do multiple regexes.

What's happening here is that the alternation operator is being used to look for the following patterns (in order):

...\\E\\        (catch potential instance of \.br\E\)
..\\E\\         (catch potential instance of \R\E\)
.\\E\\          (catch \E\ if it's at index 1 in the string)
$\\E\\          (catch \E\ if it's at index 0 in the string)

Once you've caught all of those possibilities, you can just examine what it was that was picked up before the \\E\\ . If it's \\.br or \\R , you just return the whole match (including the characters before). If it's anything else, you can return the prefix followed by a replacement for \\E\\ .

I hope this explanation makes sense!

Answer 3

Maybe this

 # /^(?=.*\\E\\)(?!.*\\\.br\\E\\)(?!.*\\R\\E\\).+$/

 ^ 
 (?= .*  \\E\\ )
 (?! .*  \\\.br\\E\\ )
 (?! .*  \\R\\E\\ )
 .+                   // can be excluded if just a test
 $                    // can be excluded if just a test

Edit if that test matches, run the global replacement for /\\\\E\\\\/g its simple and fast.