I am trying to extract from this kind of string ou=persons,ou=(.*),dc=company,dc=org the last string immediately preceded by a coma not followed by (.*). In the last case, this should give dc=company,dc=org.
Looking on regex, this seems to be a positive look behind (preceded by) of a negative look ahead.
So I have achieve this regex: (?<=(,(?!.*\Q(.*)\E))).*, but it returns ,dc=company,dc=org with the coma. I want the same thing without the coma. What I am doing wrong?
The comma appears because the capturing group contains it.
You can make the outside capture group noncapturing with (?:)
(?<=(?:,(?!.*\Q(.*)\E))).*
It seems that I have solved my problem alone, removing the capturing group around the negative look ahead. It gives the following regex: (?<=,(?!.*\Q(.*)\E)).*.
It is linked with the behavior of capturing groups in look arounds as explained here: http://www.regular-expressions.info/lookaround.html in the part Lookaround Is Atomic.
来源：https://stackoverflow.com/questions/17814421/java-regex-with-a-positive-look-behind-of-a-negative-look-ahead