How do you perform an intersection of two named regex character classes in Java?

Vote count:

0

I know how to find the intersection of two constant sets from this link to the Java tutorials.

Example:

 [0-9&&[345]]

Will produce a character class containing all of the characters which intersect the set of single digit integers ('[0-9]') and a subset of that set containing the integers 3, 4, and 5 ('[345]').

This works fine for most needs, but it requires that both sets be known at compile-time. Let's say I'm writing a lexical analyzer for mathematical expressions and want to find the intersection between the set of unary operators and binary operators. Currently, the only operator that exists in both sets is '-'. As a binary operator '-' represents subtraction but as a unary operator it also represents negation. I anticipate that other operators may be pressed into dual-usage in the future, and want to develop a solution that not only works for the present case, but also the future case without changing the code. To do that, I need to be able to determine at run-time which operators exist in both sets.

Let's say I have two character classes defined as follows:

public static final String UNARYOPS = "[!~-]";
public static final String BINARYOPS = "[|&%^*/+-]";

How would I find the intersection between these two named character sets using the regex engine in Java so that I can identify which operators need to be checked for ambiguity between binary and unary?

asked 5 mins ago

Dyndrilliac

442

How do you perform an intersection of two named regex character classes in Java?