括号内用逗号标记

Question

I'm desperately trying to tokenize a filter string which has following format:

price:gte:5.99,price:lte:10.00,model:in:[test1,test2]

The pitfall is that comma is used not only to separate filter options, but also as a separator for option lists: [test1,test2]

Applying standard regex tokenizer expression [^,]+ gives me:

price:gte:5.99
price:lte:10.00
model:in:[test1
test2]

But, obviously, I want:

price:gte:5.99
price:lte:10.00
model:in:[test1,test2]

How can I achieve that?

Answer 1

This might be an option:

(?:[^][, ]+\[[^]]+\]|[^, ]+)

Explanation

(?:                  # Non capturing group
  [^][,
]+\[[^]]+\] # Match not a comma newline ] or [, then match between []
  |                  # Or
  [^,
]+            # Match not a comma or a newline
)                    # Close non capturing group

Edit: As suggested by @anubhava added ][ to the character class so that this becomes [^][, ] to prevent unnecessary backtracking.