在Java上拆分字符串“;”，而不是“\\\\;”

Question

In Java I try try to use the String.split() method splitting a string on ";" , but not on "\\\\\\\\;" . (2 back-slashes followed by a semi-colon)

Ex: "aa;bb;cc\\\\;dd;ee\\\\;;ff" should be split into;

aa

bb

cc\\;dd

ee\\;

ff

How do I accomplish this using a regular expression?

Markus

Answer 1

Use

"aa;bb;cc\\;dd;ee\\;;ff".split("(?<!\\\\);");

(?<!...) is called a "zero-width lookbehind". In English, you're splitting on all ; characters that are NOT preceded by a double slash, without actually matching the double slash. The quadruple slash is to escape backslashes to the regex parser. The actual regular expression used in the split would then read:

(?<!\\);

Answer 2

This is called negative lookbehind and the syntax is like (?<!a)b . This matches on any b that isnt precended by an a. You would want something like:

(?<!\\\\);

Answer 3

Here a code example with . as separator:

String p = "hello.regex\\.brain\\.twister";
System.out.println( p );
for (String s : p.split( "(?<!\\\\)\\.", -1 )) {
  System.out.println( "-> "+ s );
}

Will Ouptut:

hello.regex\.brain\.twister
-> hello
-> regex\.brain.\twister