@Override
protected Triple nextInternal() throws IOException {
// next is null
if (null == scannerHeap) {
initialize();
}
while (!scannerHeap.isEmpty()) {
// get next triple
Triple next = popTriple();
while (!scannerHeap.isEmpty()) {
Triple peek = scannerHeap.peek().peek();
if (!order.comparator().match(next, peek)) {
break;
}
// remove top scanner
popTriple();
// next and peek match... merge them
next = Triple.merge(next, peek);
// keep on merging until all matches are done.
if (null == next) {
// triples and deletes canceled each other out
break;
}
}
// are we done? ignore left over deletes
// EDIT: deletes are filtered in another layer.
if (null != next /*&& !next.isDelete()*/) {
return next;
}
}
// all scanners done
return null;
}
@Override
public int run(String[] args) throws Exception {
Job job = new Job(getConf());
job.setJarByClass(PredicateCount.class);
job.setJobName("PredicateCount");
job.setOutputKeyClass(BytesWritable.class);
job.setOutputValueClass(IntWritable.class);
job.setMapperClass(Map.class);
job.setCombinerClass(Reduce.class);
job.setReducerClass(Reduce.class);
job.setInputFormatClass(TripleInputFormat.class);
job.setOutputFormatClass(SequenceFileOutputFormat.class);
int argc = 0;
TripleInputFormat.setStoreAddress(job, args[argc++]);
TripleInputFormat.setIndex(job, args[argc++]);
if ("-p".equals(args[argc])) {
argc++;
String s = args[argc++];
String p = args[argc++];
String o = args[argc++];
if ("*".equals(s)) s = null;
if ("*".equals(p)) p = null;
if ("*".equals(o)) o = null;
TripleInputFormat.setPattern(job, Triple.newPattern(s, p, o));
} else {
FileOutputFormat.setOutputPath(job, new Path(args[argc]));
}
boolean success = job.waitForCompletion(true);
return success ? 0 : 1;
}
public void map(NullWritable key, Triple value, Context context)
throws IOException, InterruptedException {
pred.set(value.getBuffer(), value.getPredicateOffset(), value.getPredicateLength());
context.write(pred, one);
}