How do I use positive and negative lookbehinds to match only if a certain pattern doesn't match before and after it?

Question

I am trying to split a string by comma, however I don't want to split if the comma is surrounded by quotes.

Here is what I am trying to do:

String s = "today,\"is, clear\",and sunny";
String[] split = s.split("(?>!\"[0-z]*),(?![0-z]*\")");

And the contents of split would be ["today", "\"is, clear\"", "and sunny"], but I'm having trouble getting the regex to work right.

It's a Java question but the regex should still be applicable. — Nick, May 09 '20 at 04:17
Your life would be simpler if you used a CSV parser library. — chrylis -cautiouslyoptimistic-, May 09 '20 at 13:12

score 1 · Accepted Answer · answered May 09 '20 at 04:24

One option would be to use a negative lookahead which asserts that a split on comma only happens when an even number of double quotes follows. Appreciate that commas inside double quotes would always have an odd number of double quotes following, assuming your quotes are always balanced.

String s = "today,\"is, clear\",and sunny";
String[] parts = s.split(",(?!.*\"(?:.*\".*\")*$)");
System.out.println(Arrays.toString(parts));

This prints:

[today, "is,  clear", and sunny]

How do I use positive and negative lookbehinds to match only if a certain pattern doesn't match before and after it?

1 Answers1