基于DFA的Java正则表达式引擎与Capture

Question

Are there any (free) regular expression engines for Java, that can compile a regular expression to a DFA, and do group capturing while matching the DFA ?

I've found dk.brics.automaton and jrexx, which both compile to DFA, but neither seems to be able to do group capture. While the other engines I've found seem to compile to NFA.

Answer 1

try this one (probably not DFA but faster than java.util) http://jregex.sourceforge.net/gstarted-advanced.html#ngroups , or this one: http://userguide.icu-project.org

according to that test: http://tusker.org/regex/regex_benchmark.html , both are fast (we all know that the benchmarks only tests what the creator of the benchmark wanted to test).

When I needed really fast DFA regex I have spawned a process that used grep ;-) (For a 6GB log file it cut my times from 10minutes to a few seconds).

Answer 2

我最近写了一篇： tree-regex 。

Answer 3

For C there is TRE and Google's RE2 libraries. TRE uses DFA, RE2 uses NFA (as far as I understand), both could subgroup matching. But I didn't see such a library for Java.

Answer 4

你可以试试Pat正则表达式库@ http://www.javaregex.com/ 。

Answer 5

dk.brics.automaton is DFA does appear to do capturing groups. I expect that feature is new in the two years since this question. Check out class AutomatonMatcher.

See http://www.brics.dk/automaton/doc/dk/brics/automaton/AutomatonMatcher.html#group(int )