org.apache.hadoop.hbase.KeyValue Java Examples

/**
 * @param fs
 * @param p
 * @param cacheConf
 * @param r
 * @param conf
 * @param indexMaintainers
 * @param viewConstants
 * @param regionInfo
 * @param regionStartKeyInHFile
 * @param splitKey
 * @throws IOException
 */
public IndexHalfStoreFileReader(final FileSystem fs, final Path p, final CacheConfig cacheConf,
        final Reference r, final Configuration conf,
        final Map<ImmutableBytesWritable, IndexMaintainer> indexMaintainers,
        final byte[][] viewConstants, final HRegionInfo regionInfo,
        final byte[] regionStartKeyInHFile, byte[] splitKey) throws IOException {
    super(fs, p, cacheConf, conf);
    this.splitkey = splitKey == null ? r.getSplitKey() : splitKey;
    // Is it top or bottom half?
    this.top = Reference.isTopFileRegion(r.getFileRegion());
    this.splitRow = CellUtil.cloneRow(KeyValue.createKeyValueFromKey(splitkey));
    this.indexMaintainers = indexMaintainers;
    this.viewConstants = viewConstants;
    this.regionInfo = regionInfo;
    this.regionStartKeyInHFile = regionStartKeyInHFile;
    this.offset = regionStartKeyInHFile.length;
}
 

@Test
public void testMap() throws Exception {
    MorphlineResultToSolrMapper resultMapper = new MorphlineResultToSolrMapper();
    resultMapper.configure(ImmutableMap.of(
        MorphlineResultToSolrMapper.MORPHLINE_FILE_PARAM, "src/test/resources/test-morphlines/extractHBaseCells.conf")
        );

    Cell kvA = new KeyValue(ROW, COLUMN_FAMILY_A, QUALIFIER_A, Bytes.toBytes(42));
    Cell kvB = new KeyValue(ROW, COLUMN_FAMILY_B, QUALIFIER_B, "dummy value".getBytes("UTF-8"));
    Result result = Result.create(Lists.newArrayList(kvA, kvB));

    Multimap expectedMap = ImmutableMultimap.of("fieldA", 42, "fieldB", "dummy value");

    resultMapper.map(result, updateWriter);
    verify(updateWriter).add(solrInputDocCaptor.capture());
    
    SolrInputDocument solrDocument = solrInputDocCaptor.getValue();
    assertEquals(expectedMap, toRecord(solrDocument).getFields());
}
 

@Override
@SuppressWarnings("rawtypes")
public void postPut(final ObserverContext e, final Put put,
		final WALEdit edit, final boolean writeToWAL) throws IOException {

	byte[][] colkey = KeyValue.parseColumn(Bytes.toBytes(inputColumn));
	if (colkey.length > 1) {
		List kvList = put.get(colkey[0], colkey[1]);
		Iterator kvl = kvList.iterator();

		while (kvl.hasNext()) {
			KeyValue kv = (KeyValue) kvl.next();
			Put indexPut = new Put(kv.getValue());
			colkey = KeyValue.parseColumn(Bytes.toBytes(indexColumn));
			indexPut.add(colkey[0], colkey[1], kv.getRow());
			table.put(indexPut);
		}
	}
}
 

/**
 * Move forward on the sub-lists set previously by seek.
 * @param key seek value (should be non-null)
 * @return true if there is at least one KV to read, false otherwise
 */
@Override
public synchronized boolean reseek(Cell key) {
  /*
   * See HBASE-4195 & HBASE-3855 & HBASE-6591 for the background on this implementation. This
   * code is executed concurrently with flush and puts, without locks. Two points must be known
   * when working on this code: 1) It's not possible to use the 'kvTail' and 'snapshot'
   * variables, as they are modified during a flush. 2) The ideal implementation for performance
   * would use the sub skip list implicitly pointed by the iterators 'kvsetIt' and 'snapshotIt'.
   * Unfortunately the Java API does not offer a method to get it. So we remember the last keys
   * we iterated to and restore the reseeked set to at least that point.
   */

  KeyValue kv = KeyValueUtil.ensureKeyValue(key);
  kvsetIt = kvsetAtCreation.tailSet(getHighest(kv, kvsetItRow)).iterator();
  return seekInSubLists();
}
 

/**
 * @param put an instance of Put
 * @param timestamp  timestamp to be used as cells version
 * @param commitTimestamp  timestamp to be used as commit timestamp
 * @throws IOException if a remote or network exception occurs.
 */
static public Put markPutAsCommitted(Put put, long timestamp, long commitTimestamp) {
    final Put tsput = new Put(put.getRow(), timestamp);
    propagateAttributes(put, tsput);

    Map<byte[], List<Cell>> kvs = put.getFamilyCellMap();
    for (List<Cell> kvl : kvs.values()) {
        for (Cell c : kvl) {
            KeyValue kv = KeyValueUtil.ensureKeyValue(c);
            Bytes.putLong(kv.getValueArray(), kv.getTimestampOffset(), timestamp);
            try {
                tsput.add(kv);
            } catch (IOException e) {
                // The existing Put has this Cell, so the cloned one
                // will never throw an IOException when it's added.
                throw new RuntimeException(e);
            }
            tsput.addColumn(CellUtil.cloneFamily(kv),
                    CellUtils.addShadowCellSuffixPrefix(CellUtil.cloneQualifier(kv), 0, CellUtil.cloneQualifier(kv).length),
                    kv.getTimestamp(),
                    Bytes.toBytes(commitTimestamp));
        }
    }

    return tsput;
}
 

@Test
public void testCreateKey() {
  byte[] row = Bytes.toBytes("myRow");
  byte[] qualifier = Bytes.toBytes("myQualifier");
  // Mimic what Storefile.createBloomKeyValue() does
  byte[] rowKey = KeyValueUtil.createFirstOnRow(row, 0, row.length, new byte[0], 0, 0, row, 0, 0).getKey();
  byte[] rowColKey = KeyValueUtil.createFirstOnRow(row, 0, row.length,
      new byte[0], 0, 0, qualifier, 0, qualifier.length).getKey();
  KeyValue rowKV = KeyValueUtil.createKeyValueFromKey(rowKey);
  KeyValue rowColKV = KeyValueUtil.createKeyValueFromKey(rowColKey);
  assertEquals(rowKV.getTimestamp(), rowColKV.getTimestamp());
  assertEquals(Bytes.toStringBinary(rowKV.getRowArray(), rowKV.getRowOffset(),
    rowKV.getRowLength()), Bytes.toStringBinary(rowColKV.getRowArray(), rowColKV.getRowOffset(),
    rowColKV.getRowLength()));
  assertEquals(0, rowKV.getQualifierLength());
}
 

@Test
public void testScanSameTimestamp() throws IOException {
  // returns only 1 of these 2 even though same timestamp
  KeyValue [] kvs = new KeyValue[] {
      create("R1", "cf", "a", 1, KeyValue.Type.Put, "dont-care"),
      create("R1", "cf", "a", 1, KeyValue.Type.Put, "dont-care"),
  };
  List<KeyValueScanner> scanners = Arrays.asList(
    new KeyValueScanner[] {new KeyValueScanFixture(CellComparator.getInstance(), kvs)});

  Scan scanSpec = new Scan().withStartRow(Bytes.toBytes("R1"));
  // this only uses maxVersions (default=1) and TimeRange (default=all)
  try (StoreScanner scan = new StoreScanner(scanSpec, scanInfo, null, scanners)) {
    List<Cell> results = new ArrayList<>();
    assertEquals(true, scan.next(results));
    assertEquals(1, results.size());
    assertEquals(1, scan.memstoreOnlyReads);
    assertEquals(kvs[0], results.get(0));
  }
}
 

public long dropSequence(Result result) throws SQLException {
    KeyValue statusKV = result.raw()[0];
    long timestamp = statusKV.getTimestamp();
    int statusCode = PDataType.INTEGER.getCodec().decodeInt(statusKV.getBuffer(), statusKV.getValueOffset(), null);
    SQLExceptionCode code = statusCode == 0 ? null : SQLExceptionCode.fromErrorCode(statusCode);
    if (code == null) {
        // Insert delete marker so that point-in-time sequences work
        insertSequenceValue(new SequenceValue(timestamp, true));
        return timestamp;
    }
    // TODO: We could have the server return the timestamps of the
    // delete markers and we could insert them here, but this seems
    // like overkill.
    // if (code == SQLExceptionCode.SEQUENCE_UNDEFINED) {
    // }
    throw new SQLExceptionInfo.Builder(code)
        .setSchemaName(key.getSchemaName())
        .setTableName(key.getSequenceName())
        .build().buildException();
}
 

@Override
public ReturnCode filterKeyValue(Cell cell) throws IOException {
  if (skipCellVersion(cell)) {
    return ReturnCode.NEXT_COL;
  }

  ReturnCode code = filter.filterKeyValue(cell);
  if (code == ReturnCode.NEXT_COL || code == ReturnCode.INCLUDE_AND_NEXT_COL) {
    // only store the reference to the keyvalue if we are returning NEXT_COL or INCLUDE_AND_NEXT_COL
    skipColumn = KeyValue.createFirstOnRow(cell.getRowArray(), cell.getRowOffset(), cell.getRowLength(),
                                             cell.getFamilyArray(), cell.getFamilyOffset(), cell.getFamilyLength(),
                                             cell.getQualifierArray(), cell.getQualifierOffset(),
                                             cell.getQualifierLength());
  } else {
    skipColumn = null;
  }
  return code;
}
 

@SuppressWarnings("deprecation")
private static void assertRow1(Iterator<KeyValue> iterator, byte[] expectedRowKey1) {
    KeyValue kv;
    assertTrue(iterator.hasNext());
    kv = iterator.next();
    assertArrayEquals(expectedRowKey1, kv.getRow());        
    assertEquals(name1, PVarchar.INSTANCE.toObject(kv.getValue()));
    assertTrue(iterator.hasNext());
    kv = iterator.next();
    assertArrayEquals(expectedRowKey1, kv.getRow());        
    assertEquals(now, PDate.INSTANCE.toObject(kv.getValue()));
    assertTrue(iterator.hasNext());
    kv = iterator.next();
    assertArrayEquals(expectedRowKey1, kv.getRow());        
    assertNull(PVarchar.INSTANCE.toObject(kv.getValue()));
}
 

/**
 * check to see if the kvs in the update match any of the passed columns. Generally, this is useful to for an index
 * codec to determine if a given update should even be indexed. This assumes that for any index, there are going to
 * small number of columns, versus the number of kvs in any one batch.
 */
public static boolean updateMatchesColumns(Collection<KeyValue> update, List<ColumnReference> columns) {
    // check to see if the kvs in the new update even match any of the columns requested
    // assuming that for any index, there are going to small number of columns, versus the number of
    // kvs in any one batch.
    boolean matches = false;
    outer: for (KeyValue kv : update) {
        for (ColumnReference ref : columns) {
            if (ref.matchesFamily(kv.getFamilyArray(), kv.getFamilyOffset(),
                kv.getFamilyLength())
                    && ref.matchesQualifier(kv.getQualifierArray(), kv.getQualifierOffset(),
                        kv.getQualifierLength())) {
                matches = true;
                // if a single column matches a single kv, we need to build a whole scanner
                break outer;
            }
        }
    }
    return matches;
}
 

protected void verifyScanAcrossSnapshot2(KeyValue kv1, KeyValue kv2)
    throws IOException {
  List<KeyValueScanner> memstorescanners = this.memstore.getScanners(mvcc.getReadPoint());
  assertEquals(2, memstorescanners.size());
  final KeyValueScanner scanner0 = memstorescanners.get(0);
  final KeyValueScanner scanner1 = memstorescanners.get(1);
  scanner0.seek(KeyValueUtil.createFirstOnRow(HConstants.EMPTY_START_ROW));
  scanner1.seek(KeyValueUtil.createFirstOnRow(HConstants.EMPTY_START_ROW));
  Cell n0 = scanner0.next();
  Cell n1 = scanner1.next();
  assertTrue(kv1.equals(n0) || kv1.equals(n1));
  assertTrue(kv2.equals(n0)
      || kv2.equals(n1)
      || kv2.equals(scanner0.next())
      || kv2.equals(scanner1.next()));
  assertNull(scanner0.next());
  assertNull(scanner1.next());
}
 

@Override
public Cell transformCell(Cell cell) throws IOException {
  // Convert Tephra deletes back into HBase deletes
  if (tx.getVisibilityLevel() == Transaction.VisibilityLevel.SNAPSHOT_ALL) {
    if (DeleteTracker.isFamilyDelete(cell)) {
      return new KeyValue(CellUtil.cloneRow(cell), CellUtil.cloneFamily(cell), null, cell.getTimestamp(),
                          KeyValue.Type.DeleteFamily);
    } else if (isColumnDelete(cell)) {
      // Note: in some cases KeyValue.Type.Delete is used in Delete object,
      // and in some other cases KeyValue.Type.DeleteColumn is used.
      // Since Tephra cannot distinguish between the two, we return KeyValue.Type.DeleteColumn.
      // KeyValue.Type.DeleteColumn makes both CellUtil.isDelete and CellUtil.isDeleteColumns return true, and will
      // work in both cases.
      return new KeyValue(CellUtil.cloneRow(cell), CellUtil.cloneFamily(cell), CellUtil.cloneQualifier(cell),
                          cell.getTimestamp(), KeyValue.Type.DeleteColumn);
    }
  }
  return cell;
}
 

private Path writeStoreFile() throws IOException {
  Path storeFileParentDir = new Path(TEST_UTIL.getDataTestDir(), "TestHFile");
  HFileContext meta = new HFileContextBuilder().withBlockSize(64 * 1024).build();
  StoreFileWriter sfw =
      new StoreFileWriter.Builder(conf, fs).withOutputDir(storeFileParentDir)
        .withFileContext(meta).build();

  final int rowLen = 32;
  Random RNG = new Random();
  for (int i = 0; i < 1000; ++i) {
    byte[] k = RandomKeyValueUtil.randomOrderedKey(RNG, i);
    byte[] v = RandomKeyValueUtil.randomValue(RNG);
    int cfLen = RNG.nextInt(k.length - rowLen + 1);
    KeyValue kv =
        new KeyValue(k, 0, rowLen, k, rowLen, cfLen, k, rowLen + cfLen,
            k.length - rowLen - cfLen, RNG.nextLong(), generateKeyType(RNG), v, 0, v.length);
    sfw.append(kv);
  }

  sfw.close();
  return sfw.getPath();
}
 

/**
 * Ensure that expired delete family markers don't override valid puts
 */
@Test
public void testExpiredDeleteFamily() throws Exception {
  long now = System.currentTimeMillis();
  KeyValue[] kvs = new KeyValue[] {
    new KeyValue(Bytes.toBytes("R1"), Bytes.toBytes("cf"), null, now-1000,
      KeyValue.Type.DeleteFamily),
    create("R1", "cf", "a", now-10, KeyValue.Type.Put,
      "dont-care"),
  };
  List<KeyValueScanner> scanners = scanFixture(kvs);
  Scan scan = new Scan();
  scan.readVersions(1);
  // scanner with ttl equal to 500
  ScanInfo scanInfo = new ScanInfo(CONF, CF, 0, 1, 500, KeepDeletedCells.FALSE,
      HConstants.DEFAULT_BLOCKSIZE, 0, CellComparator.getInstance(), false);
  try (StoreScanner scanner = new StoreScanner(scan, scanInfo, null, scanners)) {
    List<Cell> results = new ArrayList<>();
    assertEquals(true, scanner.next(results));
    assertEquals(1, results.size());
    assertEquals(kvs[1], results.get(0));
    results.clear();

    assertEquals(false, scanner.next(results));
  }
}
 

@Test
public void testCountOfCellsAfterFlatteningByScan() throws IOException {
  String[] keys1 = { "A", "B", "C" }; // A, B, C
  addRowsByKeysWith50Cols(memstore, keys1);
  // this should only flatten as there are no duplicates
  ((CompactingMemStore) memstore).flushInMemory();
  while (((CompactingMemStore) memstore).isMemStoreFlushingInMemory()) {
    Threads.sleep(10);
  }
  List<KeyValueScanner> scanners = memstore.getScanners(Long.MAX_VALUE);
  // seek
  int count = 0;
  for(int i = 0; i < scanners.size(); i++) {
    scanners.get(i).seek(KeyValue.LOWESTKEY);
    while (scanners.get(i).next() != null) {
      count++;
    }
  }
  assertEquals("the count should be ", 150, count);
  for(int i = 0; i < scanners.size(); i++) {
    scanners.get(i).close();
  }
}
 

public String call(Result result) throws Exception {
  Iterator<KeyValue> it = result.list().iterator();
  StringBuilder b = new StringBuilder();

  b.append(Bytes.toString(result.getRow()) + ":");

  while (it.hasNext()) {
    KeyValue kv = it.next();
    String q = Bytes.toString(kv.getQualifier());
    if (q.equals("counter")) {
      b.append("(" + Bytes.toString(kv.getQualifier()) + ","
          + Bytes.toLong(kv.getValue()) + ")");
    } else {
      b.append("(" + Bytes.toString(kv.getQualifier()) + ","
          + Bytes.toString(kv.getValue()) + ")");
    }
  }
  return b.toString();
}
 

@Override
public KeyValue getNextKeyHint(KeyValue currentKV) {
  //hintCount++;
  KeyValue hint = null;
      
  if (this.hintOffset >= 0 && this.hintOffset <= this.rangekeys.length - slicesLength) {
    hint = KeyValueUtil.createFirstOnRow(this.rangekeys, this.hintOffset, (short) (this.bounds[1] + 1));
    minRange = (hintOffset / this.slicesLength) / 2;
  } else {
    done = true;
  }

  /*
  byte[] row = currentKV.getRowArray();
  System.out.println("getNextKeyHint " + encodeHex(row, currentKV.getRowOffset(), currentKV.getRowLength()) + " nvalues = " + this.nvalues + " count = " + this.count + " hintOffset = " + hintOffset);
  if (null != hint) {
    row = hint.getRowArray();
    System.out.println("  hint = " + encodeHex(row, hint.getRowOffset(), hint.getRowLength())); 
  } else {
    System.out.println(" hint = null");
  }
  */
  
  return hint;
}
 

@Override
public void rollback(KeyValue kv) {
  if (LOG.isDebugEnabled()) {
    LOG.debug("Rolling back: " + toString(kv));
  }
  // If the key is in the store, delete it
  this.kvset.remove(kv);
  if (LOG.isTraceEnabled()) {
    dump();
  }
}
 

protected void addStatementInternal(Resource subj, IRI pred, Value obj, Resource context, long timestamp) throws SailException {
    if (!isWritable()) throw new SailException(tableName + " is read only");
    try {
        for (KeyValue kv : HalyardTableUtils.toKeyValues(subj, pred, obj, context, false, timestamp)) { //serialize the key value pairs relating to the statement in HBase
            put(kv);
        }
    } catch (IOException e) {
        throw new SailException(e);
    }
}
 

private long scanBlocks(int entryCount, HFileContext context, List<KeyValue> keyValues,
    FSDataInputStream fsdis, FixedFileTrailer trailer, HFileContext meta,
    HFileBlock.FSReader blockReader, int entriesRead, int blocksRead,
    DataBlockEncoder encoder) throws IOException {
  // Scan blocks the way the reader would scan them
  fsdis.seek(0);
  long curBlockPos = 0;
  while (curBlockPos <= trailer.getLastDataBlockOffset()) {
    HFileBlockDecodingContext ctx = blockReader.getBlockDecodingContext();
    HFileBlock block = blockReader.readBlockData(curBlockPos, -1, false, false, true)
      .unpack(context, blockReader);
    Assert.assertEquals(BlockType.ENCODED_DATA, block.getBlockType());
    ByteBuff origBlock = block.getBufferReadOnly();
    int pos = block.headerSize() + DataBlockEncoding.ID_SIZE;
    origBlock.position(pos);
    origBlock.limit(pos + block.getUncompressedSizeWithoutHeader() - DataBlockEncoding.ID_SIZE);
    ByteBuff buf =  origBlock.slice();
    DataBlockEncoder.EncodedSeeker seeker =
      encoder.createSeeker(encoder.newDataBlockDecodingContext(meta));
    seeker.setCurrentBuffer(buf);
    Cell res = seeker.getCell();
    KeyValue kv = keyValues.get(entriesRead);
    Assert.assertEquals(0, CellComparatorImpl.COMPARATOR.compare(res, kv));
    ++entriesRead;
    while(seeker.next()) {
      res = seeker.getCell();
      kv = keyValues.get(entriesRead);
      Assert.assertEquals(0, CellComparatorImpl.COMPARATOR.compare(res, kv));
      ++entriesRead;
    }
    ++blocksRead;
    curBlockPos += block.getOnDiskSizeWithHeader();
  }
  LOG.info("Finished reading: entries={}, blocksRead = {}", entriesRead, blocksRead);
  Assert.assertEquals(entryCount, entriesRead);
  return curBlockPos;
}
 

public KeyValue create(byte[] keyBytes, int keyOffset, int keyLength, byte[] value, int voffset, int vlen) {
    return new KeyValue(keyBytes, keyOffset, keyLength, //
            cfBytes, 0, cfBytes.length, //
            qBytes, 0, qBytes.length, //
            timestamp, KeyValue.Type.Put, //
            value, voffset, vlen);
}
 

/**
 * Sanity check that we can move logs around while we are reading
 * from them. Should this test fail, ReplicationSource would have a hard
 * time reading logs that are being archived.
 */
@Test
public void testLogMoving() throws Exception{
  Path logPath = new Path(logDir, "log");
  if (!FS.exists(logDir)) FS.mkdirs(logDir);
  if (!FS.exists(oldLogDir)) FS.mkdirs(oldLogDir);
  WALProvider.Writer writer = WALFactory.createWALWriter(FS, logPath,
      TEST_UTIL.getConfiguration());
  for(int i = 0; i < 3; i++) {
    byte[] b = Bytes.toBytes(Integer.toString(i));
    KeyValue kv = new KeyValue(b,b,b);
    WALEdit edit = new WALEdit();
    edit.add(kv);
    WALKeyImpl key = new WALKeyImpl(b, TableName.valueOf(b), 0, 0,
        HConstants.DEFAULT_CLUSTER_ID);
    writer.append(new WAL.Entry(key, edit));
    writer.sync(false);
  }
  writer.close();

  WAL.Reader reader = WALFactory.createReader(FS, logPath, TEST_UTIL.getConfiguration());
  WAL.Entry entry = reader.next();
  assertNotNull(entry);

  Path oldLogPath = new Path(oldLogDir, "log");
  FS.rename(logPath, oldLogPath);

  entry = reader.next();
  assertNotNull(entry);

  entry = reader.next();
  entry = reader.next();

  assertNull(entry);
  reader.close();
}
 

@Override
public void doReduce(RowKeyWritable row, Iterable<KeyValue> kvs, Context context) throws java.io.IOException, InterruptedException {
    for (KeyValue kv : kvs) {
        immutableBytesWritable.set(kv.getKey());
        context.write(immutableBytesWritable, kv);
    }
}
 

/** See {@link StoreFileManager#getCandidateFilesForRowKeyBefore(KeyValue)}
 * for details on this methods. */
@Override
public Iterator<HStoreFile> getCandidateFilesForRowKeyBefore(final KeyValue targetKey) {
  KeyBeforeConcatenatedLists result = new KeyBeforeConcatenatedLists();
  // Order matters for this call.
  result.addSublist(state.level0Files);
  if (!state.stripeFiles.isEmpty()) {
    int lastStripeIndex = findStripeForRow(CellUtil.cloneRow(targetKey), false);
    for (int stripeIndex = lastStripeIndex; stripeIndex >= 0; --stripeIndex) {
      result.addSublist(state.stripeFiles.get(stripeIndex));
    }
  }
  return result.iterator();
}
 

@Override
public void doMap(KEYIN key, Object value, Context context) throws IOException, InterruptedException {
    Collection<String[]> rowCollection = lookupTableInputFormat.parseMapperInput(value);
    for (String[] row : rowCollection) {
        HBaseRow hBaseRow = encoder.encode(row);

        byte[] rowKey = hBaseRow.getRowKey();
        Map<byte[], byte[]> qualifierValMap = hBaseRow.getQualifierValMap();
        outputKey.set(rowKey);
        for (Entry<byte[], byte[]> qualifierValEntry : qualifierValMap.entrySet()) {
            KeyValue outputValue = createKeyValue(rowKey, qualifierValEntry.getKey(), qualifierValEntry.getValue());
            context.write(outputKey, outputValue);
        }
    }
}
 

/**
 * Verify the data described by <code>simple_tsv</code> matches
 * <code>simple_expected</code>.
 */
protected void doLoadIncrementalHFiles(Path hfiles, TableName tableName)
    throws Exception {

  String[] args = { hfiles.toString(), tableName.getNameAsString() };
  LOG.info(format("Running LoadIncrememntalHFiles with args: %s", Arrays.asList(args)));
  assertEquals("Loading HFiles failed.", 0,
    ToolRunner.run(new BulkLoadHFilesTool(getConf()), args));

  Table table = null;
  Scan scan = new Scan() {{
    setCacheBlocks(false);
    setCaching(1000);
  }};
  try {
    table = util.getConnection().getTable(tableName);
    Iterator<Result> resultsIt = table.getScanner(scan).iterator();
    Iterator<KeyValue> expectedIt = simple_expected.iterator();
    while (resultsIt.hasNext() && expectedIt.hasNext()) {
      Result r = resultsIt.next();
      for (Cell actual : r.rawCells()) {
        assertTrue(
          "Ran out of expected values prematurely!",
          expectedIt.hasNext());
        KeyValue expected = expectedIt.next();
        assertEquals("Scan produced surprising result", 0,
          CellComparator.getInstance().compare(expected, actual));
      }
    }
    assertFalse("Did not consume all expected values.", expectedIt.hasNext());
    assertFalse("Did not consume all scan results.", resultsIt.hasNext());
  } finally {
    if (null != table) table.close();
  }
}
 

private void regexPatternFilterTests(Filter filter)
    throws Exception {
  KeyValue cell = new KeyValue(ROW, COLUMN_FAMILY, COLUMN_QUALIFIER,
    FULLSTRING_1);
  assertTrue("regexTrue",
    filter.filterCell(cell) == Filter.ReturnCode.INCLUDE);
  byte[] buffer = cell.getBuffer();
  Cell c = new ByteBufferKeyValue(ByteBuffer.wrap(buffer), 0, buffer.length);
  assertTrue("regexTrue", filter.filterCell(c) == Filter.ReturnCode.INCLUDE);
  assertFalse("regexFilterAllRemaining", filter.filterAllRemaining());
  assertFalse("regexFilterNotNull", filter.filterRow());
}
 

/**
 * Test for HBASE-8012
 */
@Test
public void testReseek() throws Exception {
  // write the file
  Path f = new Path(ROOT_DIR, name.getMethodName());
  HFileContext meta = new HFileContextBuilder().withBlockSize(8 * 1024).build();
  // Make a store file and write data to it.
  StoreFileWriter writer = new StoreFileWriter.Builder(conf, cacheConf, this.fs).withFilePath(f)
    .withFileContext(meta).build();

  writeStoreFile(writer);
  writer.close();

  ReaderContext context = new ReaderContextBuilder().withFileSystemAndPath(fs, f).build();
  HFileInfo fileInfo = new HFileInfo(context, conf);
  StoreFileReader reader =
    new StoreFileReader(context, fileInfo, cacheConf, new AtomicInteger(0), conf);
  fileInfo.initMetaAndIndex(reader.getHFileReader());

  // Now do reseek with empty KV to position to the beginning of the file

  KeyValue k = KeyValueUtil.createFirstOnRow(HConstants.EMPTY_BYTE_ARRAY);
  StoreFileScanner s = getStoreFileScanner(reader, false, false);
  s.reseek(k);

  assertNotNull("Intial reseek should position at the beginning of the file", s.peek());
}
 

/**
 * @return A Marker WALEdit that has <code>c</code> serialized as its value
 */
public static WALEdit createCompaction(final RegionInfo hri, final CompactionDescriptor c) {
  byte [] pbbytes = c.toByteArray();
  KeyValue kv = new KeyValue(getRowForRegion(hri), METAFAMILY, COMPACTION,
    EnvironmentEdgeManager.currentTime(), pbbytes);
  return new WALEdit().add(kv, METAFAMILY); //replication scope null so this won't be replicated
}