Question

I’m using Java’s matcher to group terms in a string using the following regex:

Pattern.compile("(\\\\\"[^\\\\\"]*\\\\\"|[^\\s\\\\\"]+)");

This is the part I’m having trouble with: [^\s\\\”]

I’d like it to only match non-spaces and dangling escaped quotes such as \”. Is there any way to group the \\ and \” within a character class so they’re only matched together?

I tried to use lookahead/lookbehind, but found that including it within the character class put me back at square one.

Answer 1

A character class matches a single character. If I understood you correctly, you want to match only the string \". To do this, you don’t need a character class at all–the regex \\" matches that already! (Inside a Java string, it would look like \\\\\" which is ridiculous, but there you have it.)

You can group things together using parentheses: (\\\\\"). You can also alternate inside a group like this using |. So to match non-spaces or \", you can do this: (\S|\\\\\"). (Note that \S is the same as [^\s].)

EDIT: I wasn’t paying enough attention. You can match everything but \" or a space as follows: (\\\\(?!")|[^\s\\]), I think.

How about this: ([^\\s\\\\]|\\\\(?!")). This should match anything except whitespace or \ or a \ not followed by a ".