Greedy vs lazy quantifiers in regex

Quantifiers (*, +, ?, {n,m}) are greedy by default: they match as many characters as possible, then give some back (backtrack) only if the rest of the pattern would fail otherwise. Adding ? makes them lazy: *?, +?, ??, {n,m}? match as few characters as possible and take more only when needed.

The classic example

node - <<'EOF'
const html = '<b>bold</b> and <i>italic</i>';
console.log("greedy <.+>   ", html.match(/<.+>/g));
console.log("lazy   <.+?>  ", html.match(/<.+?>/g));
console.log("class  <[^>]+>", html.match(/<[^>]+>/g));
const csv = '"a","b,c","d"';
console.log("greedy \".*\"  ", csv.match(/".*"/g));
console.log("lazy   \".*?\" ", csv.match(/".*?"/g));
EOF

greedy <.+>    [ '<b>bold</b> and <i>italic</i>' ]
lazy   <.+?>   [ '<b>', '</b>', '<i>', '</i>' ]
class  <[^>]+> [ '<b>', '</b>', '<i>', '</i>' ]
greedy ".*"   [ '"a","b,c","d"' ]
lazy   ".*?"  [ '"a"', '"b,c"', '"d"' ]

The greedy <.+> runs to the last > on the line and returns one big match. The lazy <.+?> stops at the first > it can. A negated class, <[^>]+>, gives the same result as the lazy form without backtracking, and is usually the better choice.

In short

Laziness changes which match is found, not whether one is found. For parsing real HTML, JSON or CSV, use a parser; regexes break on nesting and escaping.