private void buildNodeTableBPTreeIndex() {
try {
// Node table B+Tree index (i.e. node2id.dat/idn)
log.info(
"Node Table (3/3): building node table B+Tree index (i.e. node2id.dat and node2id.idn files)...");
final ProgressLogger monitor03 =
new ProgressLogger(
log,
"records for node table (3/3) phase",
BulkLoader.DataTickPoint,
BulkLoader.superTick);
monitor03.start();
String path = dsg.getLocation().getDirectoryPath();
new File(path, "node2id.dat").delete();
new File(path, "node2id.idn").delete();
final RecordFactory recordFactory = new RecordFactory(LenNodeHash, SizeOfNodeId);
Transform<Pair<byte[], byte[]>, Record> transformPair2Record =
new Transform<Pair<byte[], byte[]>, Record>() {
@Override
public Record convert(Pair<byte[], byte[]> pair) {
monitor03.tick();
return recordFactory.create(pair.getLeft(), pair.getRight());
}
};
int order = BPlusTreeParams.calcOrder(SystemTDB.BlockSize, recordFactory);
BPlusTreeParams bptParams = new BPlusTreeParams(order, recordFactory);
int readCacheSize = 10;
int writeCacheSize = 100;
FileSet destination = new FileSet(dsg.getLocation(), Names.indexNode2Id);
BlockMgr blkMgrNodes =
BlockMgrFactory.create(
destination, Names.bptExtTree, SystemTDB.BlockSize, readCacheSize, writeCacheSize);
BlockMgr blkMgrRecords =
BlockMgrFactory.create(
destination, Names.bptExtRecords, SystemTDB.BlockSize, readCacheSize, writeCacheSize);
Iterator<Record> iter2 = Iter.iter(sdb03.iterator()).map(transformPair2Record);
BPlusTree bpt2 =
BPlusTreeRewriter.packIntoBPlusTree(
iter2, bptParams, recordFactory, blkMgrNodes, blkMgrRecords);
bpt2.sync();
ProgressLogger.print(log, monitor03);
} finally {
sdb03.close();
sdb03 = null;
}
}
public NodeTableBuilder2(
DatasetGraphTDB dsg,
ProgressLogger monitor,
DataBag<Tuple<Long>> outputTriples,
DataBag<Tuple<Long>> outputQuads) {
// dsg.getTripleTable().getNodeTupleTable().close() ;
this.dsg = dsg;
this.monitor = monitor;
this.log = monitor.getLogger();
String filename =
new FileSet(dsg.getLocation(), Names.indexId2Node).filename(Names.extNodeData);
this.objects = FileFactory.createObjectFileDisk(filename);
this.outputTriples = outputTriples;
this.outputQuads = outputQuads;
this.stats = new StatsCollectorNodeId();
this.sdb01 =
new MultiThreadedSortedDataBag<Pair<byte[], byte[]>>(
getThresholdPolicy(), serializationFactory, new PairComparator());
try {
this.digest = MessageDigest.getInstance("MD5");
} catch (NoSuchAlgorithmException e) {
throw new AtlasException(e);
}
}