What is wrong with this regex lookahead?

Question

I have lines that start with a number followed by a space, a key and then the value. I need only the value. The problem is that my regex returns the HOLE line instead of only the value.

lines example:

1 NAME Bob /Cox/
1 SEX M
1 FAMS @F1@

regex:

/^(?=1\s[\w]*\s)[\w\W\s]*/

Did i misunderstood how "positive lookahead" works?

@CommuSoft it is a part of the value (the last name is in slashes - gedcom file format) — user3140372, Mar 17 '14 at 15:01
Lookaheads assert whether something is true or not. In this case, since all your strings follow the format `1\s[\w]*\s` and contain only `\w\W\s` characters, the string will always entirely match. — tenub, Mar 17 '14 at 15:03

score 2 · Answer 1 · answered Mar 17 '14 at 15:15

Did i misunderstood how "positive lookahead" works?

Probably. Lookahead doesn't mean "jump over this" (and don't include it in the match), but "only match if this lookahead matches from here" (and don't include it in the match, just go on from where the lookahead started).

You would need lookbehind here, which is not supported by javascript. There are a few ways around it, the easiest would be to use capturing groups here:

var r = /^1\s([\w]*)\s([\w\W\s]*)/;
lines.split("\n").forEach(function(line) {
    var m = line.match(r);
    if (m) {
        var key = m[1];
        var value = m[2];
        console.log(value);
    }
});

score 1 · Accepted Answer · answered Mar 17 '14 at 15:05

The regex you are you are using it first checks using lookahead((?=1\s[\w]*\s)) whether a certain pattern is available or not upfront. It doesn't mean that it is skipping these matches. Its only checking it's existence. Which means your next portion of regex still starts exactly after the start of string(^).

For your particular this case you can use replace instead of regex matching:

var input = '1 NAME Bob /Cox/\n1 SEX M\n1 FAMS @F1@';
input = input.replace(/^1\s[\w]*\s/mg, "");

tenub · Answer 3 · 2014-03-17T15:23:17.530

var r,s='1 NAME Bob /Cox/\n1 SEX M\n1 FAMS @F1@()',m=[],
    re=new RegExp('1 [A-Z]+ (.+)','g');
while(r=re.exec(s)){
    m.push(r[1]);
}
// m=["Bob /Cox/", "M", "@F1@()"]

Or a more robust approach:

var r,s='1 NAME Bob /Cox/\n1 SEX M\n1 FAMS @F1@()',m=[],
    re=new RegExp('1 ([A-Z]+) (.+)','g');
while(r=re.exec(s)){
    var o={};
    o[r[1]]=r[2];
    m.push(o);
}
// m=[{NAME:"Bob /Cox/"},{SEX:"M"},{FAMS:"@F1@()"}]

What is wrong with this regex lookahead?

3 Answers3