/**
* This utility method creates a list of Thrift TRowResult "struct" based on an Hbase RowResult
* object. The empty list is returned if the input is null.
*
* @param in Hbase RowResult object
* @param sortColumns This boolean dictates if row data is returned in a sorted order sortColumns
* = True will set TRowResult's sortedColumns member which is an ArrayList of TColumn struct
* sortColumns = False will set TRowResult's columns member which is a map of columnName and
* TCell struct
* @return Thrift TRowResult array
*/
public static List<TRowResult> rowResultFromHBase(Result[] in, boolean sortColumns) {
List<TRowResult> results = new ArrayList<TRowResult>();
for (Result result_ : in) {
if (result_ == null || result_.isEmpty()) {
continue;
}
TRowResult result = new TRowResult();
result.row = ByteBuffer.wrap(result_.getRow());
if (sortColumns) {
result.sortedColumns = new ArrayList<TColumn>();
for (Cell kv : result_.rawCells()) {
result.sortedColumns.add(
new TColumn(
ByteBuffer.wrap(
KeyValue.makeColumn(CellUtil.cloneFamily(kv), CellUtil.cloneQualifier(kv))),
new TCell(ByteBuffer.wrap(CellUtil.cloneValue(kv)), kv.getTimestamp())));
}
} else {
result.columns = new TreeMap<ByteBuffer, TCell>();
for (Cell kv : result_.rawCells()) {
result.columns.put(
ByteBuffer.wrap(
KeyValue.makeColumn(CellUtil.cloneFamily(kv), CellUtil.cloneQualifier(kv))),
new TCell(ByteBuffer.wrap(CellUtil.cloneValue(kv)), kv.getTimestamp()));
}
}
results.add(result);
}
return results;
}
/**
* Get the field out of the row without checking whether parsing is needed. This is called by both
* getField and getFieldsAsList.
*
* @param fieldID The id of the field starting from 0.
* @return The value of the field
*/
private Object uncheckedGetField(int fieldID) {
LazyObjectBase[] fields = getFields();
boolean[] fieldsInited = getFieldInited();
if (!fieldsInited[fieldID]) {
fieldsInited[fieldID] = true;
ColumnMapping colMap = columnsMapping[fieldID];
if (!colMap.hbaseRowKey && !colMap.hbaseTimestamp && colMap.qualifierName == null) {
// it is a column family
// primitive type for Map<Key, Value> can be stored in binary format. Pass in the
// qualifier prefix to cherry pick the qualifiers that match the prefix instead of picking
// up everything
((LazyHBaseCellMap) fields[fieldID])
.init(
result,
colMap.familyNameBytes,
colMap.binaryStorage,
colMap.qualifierPrefixBytes,
colMap.isDoPrefixCut());
return fields[fieldID].getObject();
}
if (colMap.hbaseTimestamp) {
// Get the latest timestamp of all the cells as the row timestamp
long timestamp = result.rawCells()[0].getTimestamp(); // from hbase-0.96.0
for (int i = 1; i < result.rawCells().length; i++) {
timestamp = Math.max(timestamp, result.rawCells()[i].getTimestamp());
}
LazyObjectBase lz = fields[fieldID];
if (lz instanceof LazyTimestamp) {
((LazyTimestamp) lz).getWritableObject().setTime(timestamp);
} else {
((LazyLong) lz).getWritableObject().set(timestamp);
}
return lz.getObject();
}
byte[] bytes;
if (colMap.hbaseRowKey) {
bytes = result.getRow();
} else {
// it is a column i.e. a column-family with column-qualifier
bytes = result.getValue(colMap.familyNameBytes, colMap.qualifierNameBytes);
}
if (bytes == null || isNull(oi.getNullSequence(), bytes, 0, bytes.length)) {
fields[fieldID].setNull();
} else {
ByteArrayRef ref = new ByteArrayRef();
ref.setData(bytes);
fields[fieldID].init(ref, 0, bytes.length);
}
}
return fields[fieldID].getObject();
}
public static void main(String[] args) throws IOException {
Configuration conf = HBaseConfiguration.create();
HBaseHelper helper = HBaseHelper.getHelper(conf);
helper.dropTable("testtable");
helper.createTable("testtable", "colfam1", "colfam2");
System.out.println("Adding rows to table...");
helper.fillTable("testtable", 1, 10, 10, "colfam1", "colfam2");
Connection connection = ConnectionFactory.createConnection(conf);
Table table = connection.getTable(TableName.valueOf("testtable"));
// vv SingleColumnValueFilterExample
SingleColumnValueFilter filter =
new SingleColumnValueFilter(
Bytes.toBytes("colfam1"),
Bytes.toBytes("col-5"),
CompareFilter.CompareOp.NOT_EQUAL,
new SubstringComparator("val-5"));
filter.setFilterIfMissing(true);
Scan scan = new Scan();
scan.setFilter(filter);
ResultScanner scanner = table.getScanner(scan);
// ^^ SingleColumnValueFilterExample
System.out.println("Results of scan:");
// vv SingleColumnValueFilterExample
for (Result result : scanner) {
for (Cell cell : result.rawCells()) {
System.out.println(
"Cell: "
+ cell
+ ", Value: "
+ Bytes.toString(
cell.getValueArray(), cell.getValueOffset(), cell.getValueLength()));
}
}
scanner.close();
Get get = new Get(Bytes.toBytes("row-6"));
get.setFilter(filter);
Result result = table.get(get);
System.out.println("Result of get: ");
for (Cell cell : result.rawCells()) {
System.out.println(
"Cell: "
+ cell
+ ", Value: "
+ Bytes.toString(cell.getValueArray(), cell.getValueOffset(), cell.getValueLength()));
}
// ^^ SingleColumnValueFilterExample
}
public static VesselLocation getLocation(Table VTLocation_Table, String RowKey)
throws IOException {
if (RowKey != null) {
Get get = new Get(Bytes.toBytes(RowKey));
Result result = VTLocation_Table.get(get);
VesselLocation VL = new VesselLocation();
for (Cell cell : result.rawCells()) {
String Qualifier = Bytes.toString(CellUtil.cloneQualifier(cell));
String Value = Bytes.toString(CellUtil.cloneValue(cell));
if (Qualifier.equals("coordinates")) {
VL.coordinates = Value;
} else if (Qualifier.equals("speed")) {
VL.speed = Value;
} else if (Qualifier.equals("destination")) {
VL.destination = Value;
} else if (Qualifier.equals("timestamp")) {
VL.recordtime = DateTime.parse(Value, rawformatter).getMillis();
} else if (Qualifier.equals("previouslocation")) {
VL.previouslocation = Value;
} else if (Qualifier.equals("nextlocation")) {
VL.nextlocation = Value;
}
}
return VL;
} else {
return null;
}
}
@Override
public void handleLastResult(Result lastResult) {
if (lastResult == null) {
return;
}
Cell[] rawCells = lastResult.rawCells();
Cell last = rawCells[rawCells.length - 1];
byte[] row = CellUtil.cloneRow(last);
byte[] originalRow = traceIdRowKeyDistributor.getOriginalKey(row);
long reverseStartTime =
BytesUtils.bytesToLong(originalRow, PinpointConstants.APPLICATION_NAME_MAX_LEN);
this.lastRowTimestamp = TimeUtils.recoveryTimeMillis(reverseStartTime);
byte[] qualifier = CellUtil.cloneQualifier(last);
this.lastTransactionId = TransactionIdMapper.parseVarTransactionId(qualifier, 0);
this.lastTransactionElapsed = BytesUtils.bytesToInt(qualifier, 0);
if (logger.isDebugEnabled()) {
logger.debug(
"lastRowTimestamp={}, lastTransactionId={}, lastTransactionElapsed={}",
DateUtils.longToDateStr(lastRowTimestamp),
lastTransactionId,
lastTransactionElapsed);
}
}
private static Put resultToPut(ImmutableBytesWritable key, Result result) throws IOException {
Put put = new Put(key.get());
for (Cell kv : result.rawCells()) {
put.add(kv);
}
return put;
}
@Test
public void TestMap() throws Exception {
String prefix = "0000";
final String fileName = "19691231f2cd014ea28f42788214560a21a44cef";
final String mobFilePath = prefix + fileName;
ImmutableBytesWritable r = new ImmutableBytesWritable(Bytes.toBytes("r"));
final KeyValue[] kvList = new KeyValue[1];
kvList[0] =
new KeyValue(
Bytes.toBytes("row"),
Bytes.toBytes("family"),
Bytes.toBytes("column"),
Bytes.toBytes(mobFilePath));
Result columns = mock(Result.class);
when(columns.rawCells()).thenReturn(kvList);
Configuration configuration = new Configuration(TEST_UTIL.getConfiguration());
ZooKeeperWatcher zkw = new ZooKeeperWatcher(configuration, "1", new DummyMobAbortable());
TableName tn = TableName.valueOf("testSweepMapper");
TableName lockName = MobUtils.getTableLockName(tn);
String znode = ZKUtil.joinZNode(zkw.tableLockZNode, lockName.getNameAsString());
configuration.set(SweepJob.SWEEP_JOB_ID, "1");
configuration.set(SweepJob.SWEEP_JOB_TABLE_NODE, znode);
ServerName serverName = SweepJob.getCurrentServerName(configuration);
configuration.set(SweepJob.SWEEP_JOB_SERVERNAME, serverName.toString());
TableLockManager tableLockManager =
TableLockManager.createTableLockManager(configuration, zkw, serverName);
TableLock lock = tableLockManager.writeLock(lockName, "Run sweep tool");
lock.acquire();
try {
Mapper<ImmutableBytesWritable, Result, Text, KeyValue>.Context ctx =
mock(Mapper.Context.class);
when(ctx.getConfiguration()).thenReturn(configuration);
SweepMapper map = new SweepMapper();
doAnswer(
new Answer<Void>() {
@Override
public Void answer(InvocationOnMock invocation) throws Throwable {
Text text = (Text) invocation.getArguments()[0];
KeyValue kv = (KeyValue) invocation.getArguments()[1];
assertEquals(Bytes.toString(text.getBytes(), 0, text.getLength()), fileName);
assertEquals(0, Bytes.compareTo(kv.getKey(), kvList[0].getKey()));
return null;
}
})
.when(ctx)
.write(any(Text.class), any(KeyValue.class));
map.map(r, columns, ctx);
} finally {
lock.release();
}
}
@Test
public void testStartStopRow() throws Exception {
final TableName TABLENAME1 = TableName.valueOf("testStartStopRow1");
final TableName TABLENAME2 = TableName.valueOf("testStartStopRow2");
final byte[] FAMILY = Bytes.toBytes("family");
final byte[] COLUMN1 = Bytes.toBytes("c1");
final byte[] ROW0 = Bytes.toBytesBinary("\\x01row0");
final byte[] ROW1 = Bytes.toBytesBinary("\\x01row1");
final byte[] ROW2 = Bytes.toBytesBinary("\\x01row2");
Table t1 = TEST_UTIL.createTable(TABLENAME1, FAMILY);
Table t2 = TEST_UTIL.createTable(TABLENAME2, FAMILY);
// put rows into the first table
Put p = new Put(ROW0);
p.addColumn(FAMILY, COLUMN1, COLUMN1);
t1.put(p);
p = new Put(ROW1);
p.addColumn(FAMILY, COLUMN1, COLUMN1);
t1.put(p);
p = new Put(ROW2);
p.addColumn(FAMILY, COLUMN1, COLUMN1);
t1.put(p);
CopyTable copy = new CopyTable();
assertEquals(
0,
ToolRunner.run(
new Configuration(TEST_UTIL.getConfiguration()),
copy,
new String[] {
"--new.name=" + TABLENAME2,
"--startrow=\\x01row1",
"--stoprow=\\x01row2",
TABLENAME1.getNameAsString()
}));
// verify the data was copied into table 2
// row1 exist, row0, row2 do not exist
Get g = new Get(ROW1);
Result r = t2.get(g);
assertEquals(1, r.size());
assertTrue(CellUtil.matchingQualifier(r.rawCells()[0], COLUMN1));
g = new Get(ROW0);
r = t2.get(g);
assertEquals(0, r.size());
g = new Get(ROW2);
r = t2.get(g);
assertEquals(0, r.size());
t1.close();
t2.close();
TEST_UTIL.deleteTable(TABLENAME1);
TEST_UTIL.deleteTable(TABLENAME2);
}
// Get all events with exit at last location
public static Map<Integer, VesselEvent> getAllEventsStartBeforeEndAfterBeforeLocation(
Table VTEvent_Table, String IMO_str, VesselLocation location) throws IOException {
Scan getAllEventsWithExistAtLastLocation = new Scan();
getAllEventsWithExistAtLastLocation
.setStartRow(
Bytes.toBytes(
IMO_str + LpadNum(Long.MAX_VALUE - location.recordtime, 19) + "0000000000"))
.setStopRow(Bytes.toBytes(IMO_str + LpadNum(Long.MAX_VALUE, 19) + "9999999999"))
.addColumn(details, exittime);
getAllEventsWithExistAtLastLocation.setCaching(100);
Filter ExistTimeValuefilter =
new ValueFilter(
CompareFilter.CompareOp.GREATER_OR_EQUAL,
new BinaryComparator(
Bytes.toBytes(new DateTime(location.recordtime).toString(rawformatter))));
getAllEventsWithExistAtLastLocation.setFilter(ExistTimeValuefilter);
ResultScanner Result_event = VTEvent_Table.getScanner(getAllEventsWithExistAtLastLocation);
Map<Integer, VesselEvent> events = new HashMap<Integer, VesselEvent>();
for (Result res : Result_event) {
Get get = new Get(res.getRow());
get.addColumn(details, entrytime);
get.addColumn(details, entrycoordinates);
Result result = VTEvent_Table.get(get);
String rowkey = Bytes.toString(result.getRow());
String polygonid = rowkey.substring(26);
VesselEvent VE = new VesselEvent();
VE.exittime = location.recordtime;
VE.exitcoordinates = location.coordinates;
VE.destination = location.destination;
VE.polygonid = Integer.parseInt(polygonid);
for (Cell cell : result.rawCells()) {
String Qualifier = Bytes.toString(CellUtil.cloneQualifier(cell));
String Value = Bytes.toString(CellUtil.cloneValue(cell));
if (Qualifier.equals("entertime")) {
VE.entrytime = DateTime.parse(Value, rawformatter).getMillis();
} else if (Qualifier.equals("entercoordinates")) {
VE.entrycoordinates = Value;
}
}
events.put(VE.polygonid, VE);
}
Result_event.close();
return events;
}
private void doCopyTableTest(boolean bulkload) throws Exception {
final TableName TABLENAME1 = TableName.valueOf("testCopyTable1");
final TableName TABLENAME2 = TableName.valueOf("testCopyTable2");
final byte[] FAMILY = Bytes.toBytes("family");
final byte[] COLUMN1 = Bytes.toBytes("c1");
try (Table t1 = TEST_UTIL.createTable(TABLENAME1, FAMILY);
Table t2 = TEST_UTIL.createTable(TABLENAME2, FAMILY); ) {
// put rows into the first table
for (int i = 0; i < 10; i++) {
Put p = new Put(Bytes.toBytes("row" + i));
p.addColumn(FAMILY, COLUMN1, COLUMN1);
t1.put(p);
}
CopyTable copy = new CopyTable();
int code;
if (bulkload) {
code =
ToolRunner.run(
new Configuration(TEST_UTIL.getConfiguration()),
copy,
new String[] {
"--new.name=" + TABLENAME2.getNameAsString(),
"--bulkload",
TABLENAME1.getNameAsString()
});
} else {
code =
ToolRunner.run(
new Configuration(TEST_UTIL.getConfiguration()),
copy,
new String[] {
"--new.name=" + TABLENAME2.getNameAsString(), TABLENAME1.getNameAsString()
});
}
assertEquals("copy job failed", 0, code);
// verify the data was copied into table 2
for (int i = 0; i < 10; i++) {
Get g = new Get(Bytes.toBytes("row" + i));
Result r = t2.get(g);
assertEquals(1, r.size());
assertTrue(CellUtil.matchingQualifier(r.rawCells()[0], COLUMN1));
}
} finally {
TEST_UTIL.deleteTable(TABLENAME1);
TEST_UTIL.deleteTable(TABLENAME2);
}
}
private void verifyData(Region newReg, int startRow, int numRows, byte[] qf, byte[]... families)
throws IOException {
for (int i = startRow; i < startRow + numRows; i++) {
byte[] row = Bytes.toBytes("" + i);
Get get = new Get(row);
for (byte[] family : families) {
get.addColumn(family, qf);
}
Result result = newReg.get(get);
Cell[] raw = result.rawCells();
assertEquals(families.length, result.size());
for (int j = 0; j < families.length; j++) {
assertTrue(CellUtil.matchingRow(raw[j], row));
assertTrue(CellUtil.matchingFamily(raw[j], families[j]));
assertTrue(CellUtil.matchingQualifier(raw[j], qf));
}
}
}
public static List<VesselLocation> getLocationsBetween(
Table VTLocation_Table, String imo_str, long first_timestamp, long last_timestamp)
throws IOException {
// scan
// 'cdb_vessel:vessel_location',{FILTER=>"(PrefixFilter('0000003162')"}
Scan GetExistingLocations = new Scan();
GetExistingLocations.setStartRow(
Bytes.toBytes(imo_str + LpadNum(Long.MAX_VALUE - last_timestamp, 19)))
.setStopRow(Bytes.toBytes(imo_str + LpadNum(Long.MAX_VALUE - first_timestamp + 1, 19)));
GetExistingLocations.setCaching(1000);
ResultScanner Result_ExistingLocations = VTLocation_Table.getScanner(GetExistingLocations);
List<VesselLocation> result = new ArrayList<VesselLocation>();
for (Result res : Result_ExistingLocations) {
VesselLocation VL = new VesselLocation();
for (Cell cell : res.rawCells()) {
String Qualifier = Bytes.toString(CellUtil.cloneQualifier(cell));
String Value = Bytes.toString(CellUtil.cloneValue(cell));
if (Qualifier.equals("coordinates")) {
VL.coordinates = Value;
} else if (Qualifier.equals("speed")) {
VL.speed = Value;
} else if (Qualifier.equals("destination")) {
VL.destination = Value;
} else if (Qualifier.equals("timestamp")) {
VL.recordtime = DateTime.parse(Value, rawformatter).getMillis();
} else if (Qualifier.equals("previouslocation")) {
VL.previouslocation = Value;
} else if (Qualifier.equals("nextlocation")) {
VL.nextlocation = Value;
}
}
result.add(VL);
}
Result_ExistingLocations.close();
return result;
}
@Override
public String mapRow(Result result, int rowNum) throws Exception {
if (result.isEmpty()) {
return null;
}
Cell[] rawCells = result.rawCells();
if (rawCells.length == 0) {
return null;
}
String[] ret = new String[rawCells.length];
int index = 0;
for (Cell cell : rawCells) {
ret[index++] = BytesUtils.toString(CellUtil.cloneQualifier(cell));
}
return ret[0];
}
// select data
// http://blog.csdn.net/cnweike/article/details/42920547 for more detail.
public static String select(
String tableName, String begin, String end, String colFamily, String column, String value) {
String result = "";
TableName tn = TableName.valueOf(tableName);
try {
Table table = conn.getTable(tn);
Scan scan = new Scan(Bytes.toBytes(begin), Bytes.toBytes(end));
SingleColumnValueFilter scvf =
new SingleColumnValueFilter(
Bytes.toBytes(colFamily),
Bytes.toBytes(column),
CompareFilter.CompareOp.EQUAL,
new SubstringComparator(value));
scvf.setFilterIfMissing(false);
scvf.setLatestVersionOnly(true); // OK
scan.setFilter(scvf);
ResultScanner scanner = table.getScanner(scan);
List<Map<String, String>> total = new ArrayList<Map<String, String>>();
for (Result res : scanner) {
Map<String, String> map = new HashMap<String, String>();
for (Cell cell : res.rawCells())
map.put(
Bytes.toString(CellUtil.cloneQualifier(cell)),
Bytes.toString(CellUtil.cloneValue(cell)));
total.add(map);
}
ObjectMapper mapper = new ObjectMapper();
result = mapper.writeValueAsString(total);
} catch (IOException e) {
e.printStackTrace();
}
return result;
}
/**
* Get all stocks in the same exchanger
*
* @param type
*/
public static void GetStocksSymbol(String type) {
Scan scan = new Scan();
Filter ColumnFilter =
new QualifierFilter(CompareOp.EQUAL, new BinaryComparator(Bytes.toBytes("20150806195250")));
scan.setFilter(ColumnFilter);
Hbase.setStrings("Stocks", "list");
ResultScanner results = Hbase.getAllData(scan);
BufferedWriter out = null;
File LogFile = new File(type + "Stocks");
try {
out = new BufferedWriter(new FileWriter(LogFile, false));
} catch (IOException e) {
e.printStackTrace();
}
for (Result result : results) {
for (Cell cell : result.rawCells()) {
try {
JSONObject info = new JSONObject(new String(CellUtil.cloneValue(cell)));
if (info.getString("exchange").equals(type)) {
out.write(new String(CellUtil.cloneRow(cell)) + "\r\n");
}
} catch (JSONException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}
try {
out.flush(); // 把缓存区内容压入文件
out.close();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
public void testPool() throws IOException {
PoolConfig config = new PoolConfig();
config.setMaxTotal(20);
config.setMaxIdle(5);
config.setMaxWaitMillis(1000);
config.setTestOnBorrow(true);
/* properties */
Properties props = new Properties();
props.setProperty("hbase.zookeeper.quorum", "host1,host2,host3");
props.setProperty("hbase.zookeeper.property.clientPort", "2181");
props.setProperty("hbase.master", "host1:60000");
props.setProperty("hbase.rootdir", "hdfs://host1:9000/hbase");
/* connection pool */
HbaseConnectionPool pool = new HbaseConnectionPool(config, props);
HTableInterface table = null;
HConnection conn = pool.getConnection();
table = conn.getTable(TableName.valueOf("relation"));
Get get = new Get(Bytes.toBytes("rowKey"));
Result r = table.get(get);
for (Cell cell : r.rawCells()) {
System.out.println(
"Rowkey : "
+ Bytes.toString(r.getRow())
+ " Familiy:Quilifier : "
+ Bytes.toString(CellUtil.cloneQualifier(cell))
+ " Value : "
+ Bytes.toString(CellUtil.cloneValue(cell)));
}
table.close();
System.out.println(table);
pool.returnConnection(conn);
pool.close();
}
public static VesselLocation getLocationBefore(
Table VTLocation_Table, String IMO_str, long timestamp) throws IOException {
Scan getLastLocation = new Scan();
getLastLocation.setStartRow(
Bytes.toBytes(IMO_str + LpadNum(Long.MAX_VALUE - timestamp + 1, 19)));
getLastLocation.setMaxResultSize(1);
ResultScanner Result_LastLocation = VTLocation_Table.getScanner(getLastLocation);
for (Result res : Result_LastLocation) {
VesselLocation VL = new VesselLocation();
for (Cell cell : res.rawCells()) {
String Qualifier = Bytes.toString(CellUtil.cloneQualifier(cell));
String Value = Bytes.toString(CellUtil.cloneValue(cell));
if (Qualifier.equals("coordinates")) {
VL.coordinates = Value;
} else if (Qualifier.equals("speed")) {
VL.speed = Value;
} else if (Qualifier.equals("destination")) {
VL.destination = Value;
} else if (Qualifier.equals("timestamp")) {
VL.recordtime = DateTime.parse(Value, rawformatter).getMillis();
} else if (Qualifier.equals("previouslocation")) {
VL.previouslocation = Value;
} else if (Qualifier.equals("nextlocation")) {
VL.nextlocation = Value;
}
}
Result_LastLocation.close();
return VL;
}
Result_LastLocation.close();
return null;
}
// select data
// get data where its key value between two values.
// this adapt for envinfo / measureinfo / weatherinfo.
public static String select(String tableName, String begin, String end) {
String result = "";
TableName tn = TableName.valueOf(tableName);
try {
Table table = conn.getTable(tn);
Scan scan = new Scan(Bytes.toBytes(begin), Bytes.toBytes(end));
ResultScanner scanner = table.getScanner(scan);
List<Map<String, String>> total = new ArrayList<Map<String, String>>();
for (Result res : scanner) {
Map<String, String> map = new HashMap<String, String>();
for (Cell cell : res.rawCells())
map.put(
Bytes.toString(CellUtil.cloneQualifier(cell)),
Bytes.toString(CellUtil.cloneValue(cell)));
total.add(map);
}
ObjectMapper mapper = new ObjectMapper();
result = mapper.writeValueAsString(total);
} catch (IOException e) {
e.printStackTrace();
}
return result;
}
private void doIncrementalLoadTest(boolean shouldChangeRegions) throws Exception {
util = new HBaseTestingUtility();
Configuration conf = util.getConfiguration();
byte[][] splitKeys = generateRandomSplitKeys(4);
util.startMiniCluster();
try {
HTable table = util.createTable(TABLE_NAME, FAMILIES, splitKeys);
Admin admin = table.getConnection().getAdmin();
Path testDir = util.getDataTestDirOnTestFS("testLocalMRIncrementalLoad");
assertEquals("Should start with empty table", 0, util.countRows(table));
int numRegions = -1;
try (RegionLocator r = table.getRegionLocator()) {
numRegions = r.getStartKeys().length;
}
assertEquals("Should make 5 regions", numRegions, 5);
// Generate the bulk load files
util.startMiniMapReduceCluster();
runIncrementalPELoad(conf, table.getTableDescriptor(), table.getRegionLocator(), testDir);
// This doesn't write into the table, just makes files
assertEquals("HFOF should not touch actual table", 0, util.countRows(table));
// Make sure that a directory was created for every CF
int dir = 0;
for (FileStatus f : testDir.getFileSystem(conf).listStatus(testDir)) {
for (byte[] family : FAMILIES) {
if (Bytes.toString(family).equals(f.getPath().getName())) {
++dir;
}
}
}
assertEquals("Column family not found in FS.", FAMILIES.length, dir);
// handle the split case
if (shouldChangeRegions) {
LOG.info("Changing regions in table");
admin.disableTable(table.getName());
while (util.getMiniHBaseCluster()
.getMaster()
.getAssignmentManager()
.getRegionStates()
.isRegionsInTransition()) {
Threads.sleep(200);
LOG.info("Waiting on table to finish disabling");
}
util.deleteTable(table.getName());
byte[][] newSplitKeys = generateRandomSplitKeys(14);
table = util.createTable(TABLE_NAME, FAMILIES, newSplitKeys);
while (table.getRegionLocator().getAllRegionLocations().size() != 15
|| !admin.isTableAvailable(table.getName())) {
Thread.sleep(200);
LOG.info("Waiting for new region assignment to happen");
}
}
// Perform the actual load
new LoadIncrementalHFiles(conf).doBulkLoad(testDir, table);
// Ensure data shows up
int expectedRows = NMapInputFormat.getNumMapTasks(conf) * ROWSPERSPLIT;
assertEquals(
"LoadIncrementalHFiles should put expected data in table",
expectedRows,
util.countRows(table));
Scan scan = new Scan();
ResultScanner results = table.getScanner(scan);
for (Result res : results) {
assertEquals(FAMILIES.length, res.rawCells().length);
Cell first = res.rawCells()[0];
for (Cell kv : res.rawCells()) {
assertTrue(CellUtil.matchingRow(first, kv));
assertTrue(Bytes.equals(CellUtil.cloneValue(first), CellUtil.cloneValue(kv)));
}
}
results.close();
String tableDigestBefore = util.checksumRows(table);
// Cause regions to reopen
admin.disableTable(TABLE_NAME);
while (!admin.isTableDisabled(TABLE_NAME)) {
Thread.sleep(200);
LOG.info("Waiting for table to disable");
}
admin.enableTable(TABLE_NAME);
util.waitTableAvailable(TABLE_NAME);
assertEquals(
"Data should remain after reopening of regions",
tableDigestBefore,
util.checksumRows(table));
} finally {
util.shutdownMiniMapReduceCluster();
util.shutdownMiniCluster();
}
}
private void doIncrementalLoadTest(boolean shouldChangeRegions, boolean shouldKeepLocality)
throws Exception {
util = new HBaseTestingUtility();
Configuration conf = util.getConfiguration();
conf.setBoolean(HFileOutputFormat2.LOCALITY_SENSITIVE_CONF_KEY, shouldKeepLocality);
int hostCount = 1;
int regionNum = 5;
if (shouldKeepLocality) {
// We should change host count higher than hdfs replica count when MiniHBaseCluster supports
// explicit hostnames parameter just like MiniDFSCluster does.
hostCount = 3;
regionNum = 20;
}
byte[][] splitKeys = generateRandomSplitKeys(regionNum - 1);
String[] hostnames = new String[hostCount];
for (int i = 0; i < hostCount; ++i) {
hostnames[i] = "datanode_" + i;
}
util.startMiniCluster(1, hostCount, hostnames);
Table table = util.createTable(TABLE_NAME, FAMILIES, splitKeys);
Path testDir = util.getDataTestDirOnTestFS("testLocalMRIncrementalLoad");
try (RegionLocator r = util.getConnection().getRegionLocator(TABLE_NAME);
Admin admin = util.getConnection().getAdmin(); ) {
assertEquals("Should start with empty table", 0, util.countRows(table));
int numRegions = r.getStartKeys().length;
assertEquals("Should make " + regionNum + " regions", numRegions, regionNum);
// Generate the bulk load files
runIncrementalPELoad(conf, table.getTableDescriptor(), r, testDir);
// This doesn't write into the table, just makes files
assertEquals("HFOF should not touch actual table", 0, util.countRows(table));
// Make sure that a directory was created for every CF
int dir = 0;
for (FileStatus f : testDir.getFileSystem(conf).listStatus(testDir)) {
for (byte[] family : FAMILIES) {
if (Bytes.toString(family).equals(f.getPath().getName())) {
++dir;
}
}
}
assertEquals("Column family not found in FS.", FAMILIES.length, dir);
// handle the split case
if (shouldChangeRegions) {
LOG.info("Changing regions in table");
admin.disableTable(table.getName());
while (util.getMiniHBaseCluster()
.getMaster()
.getAssignmentManager()
.getRegionStates()
.isRegionsInTransition()) {
Threads.sleep(200);
LOG.info("Waiting on table to finish disabling");
}
util.deleteTable(table.getName());
byte[][] newSplitKeys = generateRandomSplitKeys(14);
table = util.createTable(TABLE_NAME, FAMILIES, newSplitKeys);
while (util.getConnection().getRegionLocator(TABLE_NAME).getAllRegionLocations().size()
!= 15
|| !admin.isTableAvailable(table.getName())) {
Thread.sleep(200);
LOG.info("Waiting for new region assignment to happen");
}
}
// Perform the actual load
new LoadIncrementalHFiles(conf).doBulkLoad(testDir, admin, table, r);
// Ensure data shows up
int expectedRows = NMapInputFormat.getNumMapTasks(conf) * ROWSPERSPLIT;
assertEquals(
"LoadIncrementalHFiles should put expected data in table",
expectedRows,
util.countRows(table));
Scan scan = new Scan();
ResultScanner results = table.getScanner(scan);
for (Result res : results) {
assertEquals(FAMILIES.length, res.rawCells().length);
Cell first = res.rawCells()[0];
for (Cell kv : res.rawCells()) {
assertTrue(CellUtil.matchingRow(first, kv));
assertTrue(Bytes.equals(CellUtil.cloneValue(first), CellUtil.cloneValue(kv)));
}
}
results.close();
String tableDigestBefore = util.checksumRows(table);
// Check region locality
HDFSBlocksDistribution hbd = new HDFSBlocksDistribution();
for (HRegion region : util.getHBaseCluster().getRegions(TABLE_NAME)) {
hbd.add(region.getHDFSBlocksDistribution());
}
for (String hostname : hostnames) {
float locality = hbd.getBlockLocalityIndex(hostname);
LOG.info("locality of [" + hostname + "]: " + locality);
assertEquals(100, (int) (locality * 100));
}
// Cause regions to reopen
admin.disableTable(TABLE_NAME);
while (!admin.isTableDisabled(TABLE_NAME)) {
Thread.sleep(200);
LOG.info("Waiting for table to disable");
}
admin.enableTable(TABLE_NAME);
util.waitTableAvailable(TABLE_NAME);
assertEquals(
"Data should remain after reopening of regions",
tableDigestBefore,
util.checksumRows(table));
} finally {
testDir.getFileSystem(conf).delete(testDir, true);
util.deleteTable(TABLE_NAME);
util.shutdownMiniCluster();
}
}
