Examples of com.cloudera.flume.reporter.histogram.RegexGroupHistogramSink

com.cloudera.flume.reporter.histogram.RegexGroupHistogramSink

This takes a regex and a group index and generates a histogram based on the value extracted. Values that do not match are not counted. For example: the group index and regex combo of: 3, (\d+):(\d+):(\d+) for the following values: 123:456:789, abc:def:xyz, 11:22:33, 55:66:33 would result in a histogram with (value, count) : (789, 1), (33,2). NOTE: the NFA-based regex algorithm used by java.util.regex.* (and in this class) is slow and does not scale. It is fully featured but has an exponential worst case running time. This will be replaced with a faster but more memory hungry and less featured DFA-based regex algorithm. (We will lose capture groups).


      String name = m.group(1);
      // String type = m.group(2); // ignore for now.
      int grp = Integer.parseInt(m.group(3));
      Pattern regex = Pattern.compile(m.group(4));
      RegexGroupHistogramSink rghs = new RegexGroupHistogramSink(name, regex,
          grp);
      l.add(rghs);


      s = raf.readLine();
    }

Examples of com.cloudera.flume.reporter.histogram.RegexGroupHistogramSink

Related Classes of com.cloudera.flume.reporter.histogram.RegexGroupHistogramSink