Parse log with regular expression

问题:

I'm looking for kind of solution for parsing the Varnish log file. It looks like:

178.232.38.87 - - [23/May/2012:14:01:05 +0200] "GET http://static.vg.no/iphone/js/front-min.js?20120509-1 HTTP/1.1" 200 2013 "http://touch.vg.no/" "Mozilla/5.0 (Linux; U; Android 2.3.3; en-no; HTC Nexus One Build/GRI40) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1"

There can be distinguished following elements:

%h %l %u %t "%r" %s %b "%{Referer}i" "%{User-agent}i"

but I still have no idea how to do this. Simple String.split(" "); won't work.

I know regular expressions has general rules, but the most suitable would be java one.

Thanks

回答1:

I'd come up with a way to build a regular expression from chunks matching the individual fields according to their possible/expected values.

    String rexa = "(\\d+(?:\\.\\d+){3})";  // an IP address
    String rexs = "(\\S+)";                // a single token (no spaces)
    String rexdt = "\\[([^\\]]+)\\]";      // something between [ and ]
    String rexstr = "\"([^\"]*?)\"";       // a quoted string
    String rexi = "(\\d+)";                // unsigned integer

    String rex = String.join( " ", rexa, rexs, rexs, rexdt, rexstr,
                              rexi, rexi, rexstr, rexstr );

    Pattern pat = Pattern.compile( rex );
    Matcher mat = pat.matcher( h );
    if( mat.matches() ){
        for( int ig = 1; ig <= mat.groupCount(); ig++ ){
            System.out.println( mat.group( ig ) );
        }
    }

It is, of course, possible to make do with rexs in place of rexa or rexi.