Search in sources :

Example 16 with TableId

use of co.cask.cdap.data2.util.TableId in project cdap by caskdata.

the class IncrementSummingScannerTest method testMultiColumnFlushAndCompact.

@Test
public void testMultiColumnFlushAndCompact() throws Exception {
    TableId tableId = TableId.from(NamespaceId.DEFAULT.getNamespace(), "testMultiColumnFlushAndCompact");
    byte[] familyBytes = Bytes.toBytes("f");
    byte[] columnBytes = Bytes.toBytes("c");
    byte[] columnBytes2 = Bytes.toBytes("c2");
    HRegion region = createRegion(tableId, familyBytes);
    try {
        region.initialize();
        long now = 1;
        byte[] row1 = Bytes.toBytes("row1");
        byte[] row2 = Bytes.toBytes("row2");
        // Initial put to row1,c2
        Put row1P = new Put(row1);
        row1P.add(familyBytes, columnBytes2, now - 1, Bytes.toBytes(5L));
        region.put(row1P);
        // Initial put to row2,c
        Put row2P = new Put(row2);
        row2P.add(familyBytes, columnBytes, now - 1, Bytes.toBytes(10L));
        region.put(row2P);
        // Generate some increments
        long ts = now;
        for (int i = 0; i < 50; i++) {
            region.put(generateIncrementPut(familyBytes, columnBytes, row1, ts));
            region.put(generateIncrementPut(familyBytes, columnBytes, row2, ts));
            region.put(generateIncrementPut(familyBytes, columnBytes2, row1, ts));
            ts++;
        }
        // First scanner represents flush scanner
        RegionScanner scanner = new IncrementSummingScanner(region, -1, region.getScanner(new Scan().setMaxVersions()), ScanType.COMPACT_RETAIN_DELETES, now + 15, -1);
        // Second scanner is a user scan, this is to help in easy asserts
        scanner = new IncrementSummingScanner(region, -1, scanner, ScanType.USER_SCAN);
        List<Cell> results = Lists.newArrayList();
        assertTrue(scanner.next(results, 10));
        assertEquals(2, results.size());
        Cell cell = results.get(0);
        assertNotNull(cell);
        assertEquals("row1", Bytes.toString(cell.getRow()));
        assertEquals("c", Bytes.toString(cell.getQualifier()));
        assertEquals(50, Bytes.toLong(cell.getValue()));
        cell = results.get(1);
        assertNotNull(cell);
        assertEquals("row1", Bytes.toString(cell.getRow()));
        assertEquals("c2", Bytes.toString(cell.getQualifier()));
        assertEquals(55, Bytes.toLong(cell.getValue()));
        results.clear();
        assertFalse(scanner.next(results, 10));
        assertEquals(1, results.size());
        cell = results.get(0);
        assertNotNull(cell);
        assertEquals("row2", Bytes.toString(cell.getRow()));
        assertEquals(60, Bytes.toLong(cell.getValue()));
    } finally {
        region.close();
    }
}
Also used : TableId(co.cask.cdap.data2.util.TableId) HRegion(org.apache.hadoop.hbase.regionserver.HRegion) RegionScanner(org.apache.hadoop.hbase.regionserver.RegionScanner) Scan(org.apache.hadoop.hbase.client.Scan) Cell(org.apache.hadoop.hbase.Cell) Put(org.apache.hadoop.hbase.client.Put) HBase96Test(co.cask.cdap.data.hbase.HBase96Test) Test(org.junit.Test)

Example 17 with TableId

use of co.cask.cdap.data2.util.TableId in project cdap by caskdata.

the class IncrementHandlerTest method createTable.

@Override
public HTable createTable(TableId tableId) throws Exception {
    HBaseTableUtil tableUtil = new HBaseTableUtilFactory(cConf).get();
    HTableDescriptorBuilder tableDesc = tableUtil.buildHTableDescriptor(tableId);
    HColumnDescriptor columnDesc = new HColumnDescriptor(FAMILY);
    columnDesc.setMaxVersions(Integer.MAX_VALUE);
    columnDesc.setValue(IncrementHandlerState.PROPERTY_TRANSACTIONAL, "false");
    tableDesc.addFamily(columnDesc);
    tableDesc.addCoprocessor(IncrementHandler.class.getName());
    HTableDescriptor htd = tableDesc.build();
    TEST_HBASE.getHBaseAdmin().createTable(htd);
    TEST_HBASE.waitUntilTableAvailable(htd.getName(), 5000);
    return tableUtil.createHTable(conf, tableId);
}
Also used : HTableDescriptorBuilder(co.cask.cdap.data2.util.hbase.HTableDescriptorBuilder) HColumnDescriptor(org.apache.hadoop.hbase.HColumnDescriptor) HBaseTableUtilFactory(co.cask.cdap.data2.util.hbase.HBaseTableUtilFactory) HBaseTableUtil(co.cask.cdap.data2.util.hbase.HBaseTableUtil) HTableDescriptor(org.apache.hadoop.hbase.HTableDescriptor)

Example 18 with TableId

use of co.cask.cdap.data2.util.TableId in project cdap by caskdata.

the class IncrementSummingScannerTest method testIncrementScanningWithBatchAndUVB.

@Test
public void testIncrementScanningWithBatchAndUVB() throws Exception {
    TableId tableId = TableId.from(NamespaceId.DEFAULT.getNamespace(), "TestIncrementSummingScannerWithUpperVisibilityBound");
    byte[] familyBytes = Bytes.toBytes("f");
    byte[] columnBytes = Bytes.toBytes("c");
    HRegion region = createRegion(tableId, familyBytes);
    try {
        region.initialize();
        long start = 0;
        long now = start;
        long counter1 = 0;
        // adding 5 delta increments
        for (int i = 0; i < 5; i++) {
            Put p = new Put(Bytes.toBytes("r1"), now++);
            p.add(familyBytes, columnBytes, Bytes.toBytes(1L));
            p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
            region.put(p);
            counter1++;
        }
        // Also: we want different combinations of batch limit & uvbs
        for (int i = 0; i < 7; i++) {
            for (int k = 0; k < 4; k++) {
                long[] uvbs = new long[k];
                for (int l = 0; l < uvbs.length; l++) {
                    uvbs[l] = start + (k + 1) * (l + 1);
                }
                verifyCounts(region, new Scan().setMaxVersions(), new long[] { counter1 }, i > 0 ? i : -1, uvbs);
            }
        }
        // Now test same with two groups of increments
        int counter2 = 0;
        for (int i = 0; i < 5; i++) {
            Put p = new Put(Bytes.toBytes("r2"), now + i);
            p.add(familyBytes, columnBytes, Bytes.toBytes(2L));
            p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
            region.put(p);
            counter2 += 2;
        }
        for (int i = 0; i < 12; i++) {
            for (int k = 0; k < 4; k++) {
                long[] uvbs = new long[k];
                for (int l = 0; l < uvbs.length; l++) {
                    uvbs[l] = start + (k + 1) * (l + 1);
                }
                verifyCounts(region, new Scan().setMaxVersions(), new long[] { counter1, counter2 }, i > 0 ? i : -1, uvbs);
            }
        }
    } finally {
        region.close();
    }
}
Also used : TableId(co.cask.cdap.data2.util.TableId) HRegion(org.apache.hadoop.hbase.regionserver.HRegion) Scan(org.apache.hadoop.hbase.client.Scan) Put(org.apache.hadoop.hbase.client.Put) HBase96Test(co.cask.cdap.data.hbase.HBase96Test) Test(org.junit.Test)

Example 19 with TableId

use of co.cask.cdap.data2.util.TableId in project cdap by caskdata.

the class IncrementSummingScannerTest method testFlushAndCompact.

@Test
public void testFlushAndCompact() throws Exception {
    TableId tableId = TableId.from(NamespaceId.DEFAULT.getNamespace(), "TestFlushAndCompact");
    byte[] familyBytes = Bytes.toBytes("f");
    byte[] columnBytes = Bytes.toBytes("c");
    HRegion region = createRegion(tableId, familyBytes);
    try {
        region.initialize();
        // load an initial set of increments
        long ts = System.currentTimeMillis();
        byte[] row1 = Bytes.toBytes("row1");
        for (int i = 0; i < 50; i++) {
            Put p = new Put(row1);
            p.add(familyBytes, columnBytes, ts, Bytes.toBytes(1L));
            p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
            ts++;
            region.put(p);
        }
        byte[] row2 = Bytes.toBytes("row2");
        ts = System.currentTimeMillis();
        // start with a full put
        Put row2P = new Put(row2);
        row2P.add(familyBytes, columnBytes, ts++, Bytes.toBytes(10L));
        region.put(row2P);
        for (int i = 0; i < 10; i++) {
            Put p = new Put(row2);
            p.add(familyBytes, columnBytes, ts++, Bytes.toBytes(1L));
            p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
            region.put(p);
        }
        // force a region flush
        region.flushcache();
        region.waitForFlushesAndCompactions();
        Result r1 = region.get(new Get(row1));
        assertNotNull(r1);
        assertFalse(r1.isEmpty());
        // row1 should have a full put aggregating all 50 incrments
        Cell r1Cell = r1.getColumnLatestCell(familyBytes, columnBytes);
        assertNotNull(r1Cell);
        assertEquals(50L, Bytes.toLong(r1Cell.getValue()));
        Result r2 = region.get(new Get(row2));
        assertNotNull(r2);
        assertFalse(r2.isEmpty());
        // row2 should have a full put aggregating prior put + 10 increments
        Cell r2Cell = r2.getColumnLatestCell(familyBytes, columnBytes);
        assertNotNull(r2Cell);
        assertEquals(20L, Bytes.toLong(r2Cell.getValue()));
        // add 30 more increments to row2
        for (int i = 0; i < 30; i++) {
            Put p = new Put(row2);
            p.add(familyBytes, columnBytes, ts++, Bytes.toBytes(1L));
            p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
            region.put(p);
        }
        // row2 should now have a full put aggregating prior 20 value + 30 increments
        r2 = region.get(new Get(row2));
        assertNotNull(r2);
        assertFalse(r2.isEmpty());
        r2Cell = r2.getColumnLatestCell(familyBytes, columnBytes);
        assertNotNull(r2Cell);
        assertEquals(50L, Bytes.toLong(r2Cell.getValue()));
        // force another region flush
        region.flushcache();
        region.waitForFlushesAndCompactions();
        // add 100 more increments to row2
        for (int i = 0; i < 100; i++) {
            Put p = new Put(row2);
            p.add(familyBytes, columnBytes, ts++, Bytes.toBytes(1L));
            p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
            region.put(p);
        }
        // row2 should now have a full put aggregating prior 50 value + 100 increments
        r2 = region.get(new Get(row2));
        assertNotNull(r2);
        assertFalse(r2.isEmpty());
        r2Cell = r2.getColumnLatestCell(familyBytes, columnBytes);
        assertNotNull(r2Cell);
        assertEquals(150L, Bytes.toLong(r2Cell.getValue()));
    } finally {
        region.close();
    }
}
Also used : TableId(co.cask.cdap.data2.util.TableId) HRegion(org.apache.hadoop.hbase.regionserver.HRegion) Get(org.apache.hadoop.hbase.client.Get) Cell(org.apache.hadoop.hbase.Cell) Put(org.apache.hadoop.hbase.client.Put) Result(org.apache.hadoop.hbase.client.Result) HBase96Test(co.cask.cdap.data.hbase.HBase96Test) Test(org.junit.Test)

Example 20 with TableId

use of co.cask.cdap.data2.util.TableId in project cdap by caskdata.

the class IncrementSummingScannerTest method testIncrementScanning.

@Test
public void testIncrementScanning() throws Exception {
    TableId tableId = TableId.from(NamespaceId.DEFAULT.getNamespace(), "TestIncrementSummingScanner");
    byte[] familyBytes = Bytes.toBytes("f");
    byte[] columnBytes = Bytes.toBytes("c");
    HRegion region = createRegion(tableId, familyBytes);
    try {
        region.initialize();
        // test handling of a single increment value alone
        Put p = new Put(Bytes.toBytes("r1"));
        p.add(familyBytes, columnBytes, Bytes.toBytes(3L));
        p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
        region.put(p);
        Scan scan = new Scan();
        RegionScanner scanner = new IncrementSummingScanner(region, -1, region.getScanner(scan), ScanType.USER_SCAN);
        List<Cell> results = Lists.newArrayList();
        scanner.next(results);
        assertEquals(1, results.size());
        Cell cell = results.get(0);
        assertNotNull(cell);
        assertEquals(3L, Bytes.toLong(cell.getValue()));
        // test handling of a single total sum
        p = new Put(Bytes.toBytes("r2"));
        p.add(familyBytes, columnBytes, Bytes.toBytes(5L));
        region.put(p);
        scan = new Scan(Bytes.toBytes("r2"));
        scanner = new IncrementSummingScanner(region, -1, region.getScanner(scan), ScanType.USER_SCAN);
        results = Lists.newArrayList();
        scanner.next(results);
        assertEquals(1, results.size());
        cell = results.get(0);
        assertNotNull(cell);
        assertEquals(5L, Bytes.toLong(cell.getValue()));
        // test handling of multiple increment values
        long now = System.currentTimeMillis();
        p = new Put(Bytes.toBytes("r3"));
        for (int i = 0; i < 5; i++) {
            p.add(familyBytes, columnBytes, now - i, Bytes.toBytes((long) (i + 1)));
        }
        p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
        region.put(p);
        scan = new Scan(Bytes.toBytes("r3"));
        scan.setMaxVersions();
        scanner = new IncrementSummingScanner(region, -1, region.getScanner(scan), ScanType.USER_SCAN);
        results = Lists.newArrayList();
        scanner.next(results);
        assertEquals(1, results.size());
        cell = results.get(0);
        assertNotNull(cell);
        assertEquals(15L, Bytes.toLong(cell.getValue()));
        // test handling of multiple increment values followed by a total sum, then other increments
        now = System.currentTimeMillis();
        p = new Put(Bytes.toBytes("r4"));
        for (int i = 0; i < 3; i++) {
            p.add(familyBytes, columnBytes, now - i, Bytes.toBytes(1L));
        }
        p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
        region.put(p);
        // this put will appear as a "total" sum prior to all the delta puts
        p = new Put(Bytes.toBytes("r4"));
        p.add(familyBytes, columnBytes, now - 5, Bytes.toBytes(5L));
        region.put(p);
        scan = new Scan(Bytes.toBytes("r4"));
        scan.setMaxVersions();
        scanner = new IncrementSummingScanner(region, -1, region.getScanner(scan), ScanType.USER_SCAN);
        results = Lists.newArrayList();
        scanner.next(results);
        assertEquals(1, results.size());
        cell = results.get(0);
        assertNotNull(cell);
        assertEquals(8L, Bytes.toLong(cell.getValue()));
        // test handling of an increment column followed by a non-increment column
        p = new Put(Bytes.toBytes("r4"));
        p.add(familyBytes, Bytes.toBytes("c2"), Bytes.toBytes("value"));
        region.put(p);
        scan = new Scan(Bytes.toBytes("r4"));
        scan.setMaxVersions();
        scanner = new IncrementSummingScanner(region, -1, region.getScanner(scan), ScanType.USER_SCAN);
        results = Lists.newArrayList();
        scanner.next(results);
        assertEquals(2, results.size());
        cell = results.get(0);
        assertNotNull(cell);
        assertEquals(8L, Bytes.toLong(cell.getValue()));
        cell = results.get(1);
        assertNotNull(cell);
        assertEquals("value", Bytes.toString(cell.getValue()));
        // test handling of an increment column followed by a delete
        now = System.currentTimeMillis();
        Delete d = new Delete(Bytes.toBytes("r5"));
        d.deleteColumn(familyBytes, columnBytes, now - 3);
        region.delete(d);
        p = new Put(Bytes.toBytes("r5"));
        for (int i = 2; i >= 0; i--) {
            p.add(familyBytes, columnBytes, now - i, Bytes.toBytes(1L));
        }
        p.setAttribute(HBaseTable.DELTA_WRITE, TRUE);
        region.put(p);
        scan = new Scan(Bytes.toBytes("r5"));
        scan.setMaxVersions();
        scan.setRaw(true);
        scanner = new IncrementSummingScanner(region, -1, region.getScanner(scan), ScanType.COMPACT_RETAIN_DELETES);
        results = Lists.newArrayList();
        scanner.next(results);
        // delete marker will not be returned for user scan
        assertEquals(2, results.size());
        cell = results.get(0);
        assertNotNull(cell);
        assertEquals(3L, Bytes.toLong(cell.getValue(), IncrementHandlerState.DELTA_MAGIC_PREFIX.length, 8));
        // next cell should be the delete
        cell = results.get(1);
        assertTrue(CellUtil.isDelete(cell));
    } finally {
        region.close();
    }
}
Also used : TableId(co.cask.cdap.data2.util.TableId) Delete(org.apache.hadoop.hbase.client.Delete) HRegion(org.apache.hadoop.hbase.regionserver.HRegion) RegionScanner(org.apache.hadoop.hbase.regionserver.RegionScanner) Scan(org.apache.hadoop.hbase.client.Scan) Cell(org.apache.hadoop.hbase.Cell) Put(org.apache.hadoop.hbase.client.Put) HBase96Test(co.cask.cdap.data.hbase.HBase96Test) Test(org.junit.Test)

Aggregations

TableId (co.cask.cdap.data2.util.TableId)102 Test (org.junit.Test)49 HTableDescriptor (org.apache.hadoop.hbase.HTableDescriptor)41 HRegion (org.apache.hadoop.hbase.regionserver.HRegion)39 Put (org.apache.hadoop.hbase.client.Put)34 NamespaceId (co.cask.cdap.proto.id.NamespaceId)26 Cell (org.apache.hadoop.hbase.Cell)24 Scan (org.apache.hadoop.hbase.client.Scan)23 HTableDescriptorBuilder (co.cask.cdap.data2.util.hbase.HTableDescriptorBuilder)20 RegionScanner (org.apache.hadoop.hbase.regionserver.RegionScanner)18 HBaseTableUtil (co.cask.cdap.data2.util.hbase.HBaseTableUtil)16 HBaseTableUtilFactory (co.cask.cdap.data2.util.hbase.HBaseTableUtilFactory)15 IOException (java.io.IOException)14 HTable (org.apache.hadoop.hbase.client.HTable)12 HBaseDDLExecutor (co.cask.cdap.spi.hbase.HBaseDDLExecutor)11 HBaseAdmin (org.apache.hadoop.hbase.client.HBaseAdmin)10 Delete (org.apache.hadoop.hbase.client.Delete)9 Result (org.apache.hadoop.hbase.client.Result)9 Path (org.apache.hadoop.fs.Path)8 HColumnDescriptor (org.apache.hadoop.hbase.HColumnDescriptor)8