parsing log4j log file using regular expression

问题

I have created a java application for parsing the log4j log file using regular expression, The application is working fine for the log which i have shown below

1999-11-27 15:49:37,459 [thread-x] ERROR mypackage - Catastrophic system failure

but not working for

2015-01-22 01:52:54,237 [http-bio-80-exec-5] FATAL   TestLog4jServlet - Show FATAL message

My log4j ConversionPattern is given below

log4j.appender.Appender2.layout.ConversionPattern=%d [%t] %-7p %10c{1} - %m%n

Can anyone please tell me some solution for this

My code is as given below

public static void main(String[] args) {
    String regex = "(\\d{4}-\\d{2}-\\d{2}) (\\d{2}:\\d{2}:\\d{2},\\d{3}) \\[(.*)\\] ([^ ]*) ([^ ]*) - (.*)$";

    Pattern p = Pattern.compile(regex);
    String[] samples = {
            "2015-01-22 01:52:54,237 [http-bio-80-exec-5] FATAL   TestLog4jServlet - Show FATAL message"
        };

    Matcher m = p.matcher(samples[1]);
    System.out.println(m.matches());
    if (m.matches() && m.groupCount() == 6) {
        String date = m.group(1);
        String time = m.group(2);
        String threadId = m.group(3);
        String priority = m.group(4);
        String category = m.group(5);
        String message = m.group(6);

        System.out.println("date: " + date);
        System.out.println("time: " + time);
        System.out.println("threadId: " + threadId);
        System.out.println("priority: " + priority);
        System.out.println("category: " + category);
        System.out.println("message: " + message);
    }
}

回答1:

Because there are two spaces between FATAL and TestLog4jServlet but you included only one space in your regex. So i suggest you to replace the corresponding space with <space>+ which allows one or more spaces.

(\d{4}-\d{2}-\d{2}) (\d{2}:\d{2}:\d{2},\d{3}) \[(.*?)\] ([^ ]*) +([^ ]*) - (.*)$
                                                                ^
                                                                |

DEMO

Java regex would be,

"(\\d{4}-\\d{2}-\\d{2}) (\\d{2}:\\d{2}:\\d{2},\\d{3}) \\[(.*)\\] ([^ ]*) +([^ ]*) - (.*)$"

回答2:

I think the Logstash is better for parsing log.

来源：https://stackoverflow.com/questions/28083827/parsing-log4j-log-file-using-regular-expression

标签

java

regex

log4j