org.apache.kylin.cube.model.HBaseColumnFamilyDesc Java Examples

@Override
protected void doSetup(Context context) throws IOException {
    super.bindCurrentConfiguration(context.getConfiguration());
    cubeName = context.getConfiguration().get(BatchConstants.CFG_CUBE_NAME);

    KylinConfig config = AbstractHadoopJob.loadKylinPropsAndMetadata();

    CubeManager cubeMgr = CubeManager.getInstance(config);
    cubeDesc = cubeMgr.getCube(cubeName).getDescriptor();

    inputCodec = new MeasureCodec(cubeDesc.getMeasures());
    inputMeasures = new Object[cubeDesc.getMeasures().size()];
    keyValueCreators = Lists.newArrayList();

    for (HBaseColumnFamilyDesc cfDesc : cubeDesc.getHbaseMapping().getColumnFamily()) {
        for (HBaseColumnDesc colDesc : cfDesc.getColumns()) {
            keyValueCreators.add(new KeyValueCreator(cubeDesc, colDesc));
        }
    }
}
 

/**
 * prune untouched hbase columns
 */
protected List<Pair<byte[], byte[]>> makeHBaseColumns(ImmutableBitSet selectedColBlocks) {
    List<Pair<byte[], byte[]>> result = Lists.newArrayList();

    int colBlkIndex = 1;
    HBaseMappingDesc hbaseMapping = cubeSeg.getCubeDesc().getHbaseMapping();
    for (HBaseColumnFamilyDesc familyDesc : hbaseMapping.getColumnFamily()) {
        byte[] byteFamily = Bytes.toBytes(familyDesc.getName());
        for (HBaseColumnDesc hbaseColDesc : familyDesc.getColumns()) {
            if (selectedColBlocks.get(colBlkIndex)) {
                byte[] byteQualifier = Bytes.toBytes(hbaseColDesc.getQualifier());
                result.add(Pair.newPair(byteFamily, byteQualifier));
            }
            colBlkIndex++;
        }
    }

    return result;
}
 

/**
 * for each selected hbase column, it might contain values of multiple GT columns.
 * The mapping should be passed down to storage
 */
protected List<List<Integer>> getHBaseColumnsGTMapping(ImmutableBitSet selectedColBlocks) {

    List<List<Integer>> ret = Lists.newArrayList();

    int colBlkIndex = 1;
    int metricOffset = fullGTInfo.getPrimaryKey().trueBitCount();

    HBaseMappingDesc hbaseMapping = cubeSeg.getCubeDesc().getHbaseMapping();
    for (HBaseColumnFamilyDesc familyDesc : hbaseMapping.getColumnFamily()) {
        for (HBaseColumnDesc hbaseColDesc : familyDesc.getColumns()) {
            if (selectedColBlocks.get(colBlkIndex)) {
                int[] metricIndexes = hbaseColDesc.getMeasureIndex();
                Integer[] gtIndexes = new Integer[metricIndexes.length];
                for (int i = 0; i < gtIndexes.length; i++) {
                    gtIndexes[i] = metricIndexes[i] + metricOffset;
                }
                ret.add(Arrays.asList(gtIndexes));
            }
            colBlkIndex++;
        }
    }

    Preconditions.checkState(selectedColBlocks.trueBitCount() == ret.size() + 1);
    return ret;
}
 

@Override
protected void doSetup(Context context) throws IOException {
    super.bindCurrentConfiguration(context.getConfiguration());
    cubeName = context.getConfiguration().get(BatchConstants.CFG_CUBE_NAME);

    KylinConfig config = AbstractHadoopJob.loadKylinPropsAndMetadata();

    CubeManager cubeMgr = CubeManager.getInstance(config);
    cubeDesc = cubeMgr.getCube(cubeName).getDescriptor();

    inputCodec = new MeasureCodec(cubeDesc.getMeasures());
    inputMeasures = new Object[cubeDesc.getMeasures().size()];
    keyValueCreators = Lists.newArrayList();

    for (HBaseColumnFamilyDesc cfDesc : cubeDesc.getHbaseMapping().getColumnFamily()) {
        for (HBaseColumnDesc colDesc : cfDesc.getColumns()) {
            keyValueCreators.add(new KeyValueCreator(cubeDesc, colDesc));
        }
    }
}
 

/**
 * prune untouched hbase columns
 */
protected List<Pair<byte[], byte[]>> makeHBaseColumns(ImmutableBitSet selectedColBlocks) {
    List<Pair<byte[], byte[]>> result = Lists.newArrayList();

    int colBlkIndex = 1;
    HBaseMappingDesc hbaseMapping = cubeSeg.getCubeDesc().getHbaseMapping();
    for (HBaseColumnFamilyDesc familyDesc : hbaseMapping.getColumnFamily()) {
        byte[] byteFamily = Bytes.toBytes(familyDesc.getName());
        for (HBaseColumnDesc hbaseColDesc : familyDesc.getColumns()) {
            if (selectedColBlocks.get(colBlkIndex)) {
                byte[] byteQualifier = Bytes.toBytes(hbaseColDesc.getQualifier());
                result.add(Pair.newPair(byteFamily, byteQualifier));
            }
            colBlkIndex++;
        }
    }

    return result;
}
 

/**
 * for each selected hbase column, it might contain values of multiple GT columns.
 * The mapping should be passed down to storage
 */
protected List<List<Integer>> getHBaseColumnsGTMapping(ImmutableBitSet selectedColBlocks) {

    List<List<Integer>> ret = Lists.newArrayList();

    int colBlkIndex = 1;
    int metricOffset = fullGTInfo.getPrimaryKey().trueBitCount();

    HBaseMappingDesc hbaseMapping = cubeSeg.getCubeDesc().getHbaseMapping();
    for (HBaseColumnFamilyDesc familyDesc : hbaseMapping.getColumnFamily()) {
        for (HBaseColumnDesc hbaseColDesc : familyDesc.getColumns()) {
            if (selectedColBlocks.get(colBlkIndex)) {
                int[] metricIndexes = hbaseColDesc.getMeasureIndex();
                Integer[] gtIndexes = new Integer[metricIndexes.length];
                for (int i = 0; i < gtIndexes.length; i++) {
                    gtIndexes[i] = metricIndexes[i] + metricOffset;
                }
                ret.add(Arrays.asList(gtIndexes));
            }
            colBlkIndex++;
        }
    }

    Preconditions.checkState(selectedColBlocks.trueBitCount() == ret.size() + 1);
    return ret;
}
 

@Override
protected void setup(Context context) throws IOException {
    super.publishConfiguration(context.getConfiguration());
    cubeName = context.getConfiguration().get(BatchConstants.CFG_CUBE_NAME);

    KylinConfig config = AbstractHadoopJob.loadKylinPropsAndMetadata(context.getConfiguration());

    CubeManager cubeMgr = CubeManager.getInstance(config);
    cubeDesc = cubeMgr.getCube(cubeName).getDescriptor();

    inputCodec = new MeasureCodec(cubeDesc.getMeasures());
    inputMeasures = new Object[cubeDesc.getMeasures().size()];
    keyValueCreators = Lists.newArrayList();

    for (HBaseColumnFamilyDesc cfDesc : cubeDesc.getHBaseMapping().getColumnFamily()) {
        for (HBaseColumnDesc colDesc : cfDesc.getColumns()) {
            keyValueCreators.add(new KeyValueCreator(cubeDesc, colDesc));
        }
    }
}
 

private void initMeasureReferenceToColumnFamily() {
    if (measures == null || measures.size() == 0)
        return;

    Map<String, MeasureDesc> measureCache = new HashMap<String, MeasureDesc>();
    for (MeasureDesc m : measures)
        measureCache.put(m.getName(), m);

    for (HBaseColumnFamilyDesc cf : getHBaseMapping().getColumnFamily()) {
        for (HBaseColumnDesc c : cf.getColumns()) {
            MeasureDesc[] measureDescs = new MeasureDesc[c.getMeasureRefs().length];
            for (int i = 0; i < c.getMeasureRefs().length; i++) {
                measureDescs[i] = measureCache.get(c.getMeasureRefs()[i]);
            }
            c.setMeasures(measureDescs);
            c.setColumnFamilyName(cf.getName());
        }
    }
}
 

private void validateColumnFamily(CubeDesc cubeDesc) {
    Set<String> columnFamilyMetricsSet = Sets.newHashSet();
    for (HBaseColumnFamilyDesc hBaseColumnFamilyDesc : cubeDesc.getHbaseMapping().getColumnFamily()) {
        for (HBaseColumnDesc hBaseColumnDesc : hBaseColumnFamilyDesc.getColumns()) {
            for (String columnName : hBaseColumnDesc.getMeasureRefs()) {
                columnFamilyMetricsSet.add(columnName);
            }
        }
    }
    for (MeasureDesc measureDesc : cubeDesc.getMeasures()) {
        if (!columnFamilyMetricsSet.contains(measureDesc.getName())) {
            throw new BadRequestException("column family lack measure:" + measureDesc.getName());
        }
    }
    if (cubeDesc.getMeasures().size() != columnFamilyMetricsSet.size()) {
        throw new BadRequestException(
                "the number of input measure and the number of measure defined in cubedesc are not consistent");
    }

    for (RowKeyColDesc rowKeyColDesc : cubeDesc.getRowkey().getRowKeyColumns()) {
        Object[] encodingConf = DimensionEncoding.parseEncodingConf(rowKeyColDesc.getEncoding());
        String encodingName = (String) encodingConf[0];
        String[] encodingArgs = (String[]) encodingConf[1];

        if (!DimensionEncodingFactory.isValidEncoding(encodingName, encodingArgs,
                rowKeyColDesc.getEncodingVersion())) {
            throw new BadRequestException("Illegal row key column desc: " + rowKeyColDesc);
        }
    }
}
 

private void validateColumnFamily(CubeDesc cubeDesc) {
    Set<String> columnFamilyMetricsSet = Sets.newHashSet();
    for (HBaseColumnFamilyDesc hBaseColumnFamilyDesc : cubeDesc.getHbaseMapping().getColumnFamily()) {
        for (HBaseColumnDesc hBaseColumnDesc : hBaseColumnFamilyDesc.getColumns()) {
            for (String columnName : hBaseColumnDesc.getMeasureRefs()) {
                columnFamilyMetricsSet.add(columnName);
            }
        }
    }
    for (MeasureDesc measureDesc : cubeDesc.getMeasures()) {
        if (!columnFamilyMetricsSet.contains(measureDesc.getName())) {
            throw new BadRequestException("column family lack measure:" + measureDesc.getName());
        }
    }
    if (cubeDesc.getMeasures().size() != columnFamilyMetricsSet.size()) {
        throw new BadRequestException(
                "the number of input measure and the number of measure defined in cubedesc are not consistent");
    }

    for (RowKeyColDesc rowKeyColDesc : cubeDesc.getRowkey().getRowKeyColumns()) {
        Object[] encodingConf = DimensionEncoding.parseEncodingConf(rowKeyColDesc.getEncoding());
        String encodingName = (String) encodingConf[0];
        String[] encodingArgs = (String[]) encodingConf[1];

        if (!DimensionEncodingFactory.isValidEncoding(encodingName, encodingArgs,
                rowKeyColDesc.getEncodingVersion())) {
            throw new BadRequestException("Illegal row key column desc: " + rowKeyColDesc);
        }
    }
}
 

public static void createHTable(CubeSegment cubeSegment, byte[][] splitKeys) throws IOException {
    String tableName = cubeSegment.getStorageLocationIdentifier();
    CubeInstance cubeInstance = cubeSegment.getCubeInstance();
    CubeDesc cubeDesc = cubeInstance.getDescriptor();
    KylinConfig kylinConfig = cubeDesc.getConfig();

    HTableDescriptor tableDesc = new HTableDescriptor(TableName.valueOf(cubeSegment.getStorageLocationIdentifier()));
    tableDesc.setValue(HTableDescriptor.SPLIT_POLICY, DisabledRegionSplitPolicy.class.getName());
    tableDesc.setValue(IRealizationConstants.HTableTag, kylinConfig.getMetadataUrlPrefix());
    tableDesc.setValue(IRealizationConstants.HTableCreationTime, String.valueOf(System.currentTimeMillis()));

    if (!StringUtils.isEmpty(kylinConfig.getKylinOwner())) {
        //HTableOwner is the team that provides kylin service
        tableDesc.setValue(IRealizationConstants.HTableOwner, kylinConfig.getKylinOwner());
    }

    String commitInfo = KylinVersion.getGitCommitInfo();
    if (!StringUtils.isEmpty(commitInfo)) {
        tableDesc.setValue(IRealizationConstants.HTableGitTag, commitInfo);
    }

    //HTableUser is the cube owner, which will be the "user"
    tableDesc.setValue(IRealizationConstants.HTableUser, cubeInstance.getOwner());

    tableDesc.setValue(IRealizationConstants.HTableSegmentTag, cubeSegment.toString());

    Configuration conf = HBaseConnection.getCurrentHBaseConfiguration();
    Connection conn = HBaseConnection.get(kylinConfig.getStorageUrl());
    Admin admin = conn.getAdmin();

    try {
        if (User.isHBaseSecurityEnabled(conf)) {
            // add coprocessor for bulk load
            tableDesc.addCoprocessor("org.apache.hadoop.hbase.security.access.SecureBulkLoadEndpoint");
        }

        for (HBaseColumnFamilyDesc cfDesc : cubeDesc.getHbaseMapping().getColumnFamily()) {
            HColumnDescriptor cf = createColumnFamily(kylinConfig, cfDesc.getName(), cfDesc.isMemoryHungry());
            tableDesc.addFamily(cf);
        }

        if (admin.tableExists(TableName.valueOf(tableName))) {
            // admin.disableTable(tableName);
            // admin.deleteTable(tableName);
            throw new RuntimeException("HBase table " + tableName + " exists!");
        }

        DeployCoprocessorCLI.deployCoprocessor(tableDesc);

        admin.createTable(tableDesc, splitKeys);
        Preconditions.checkArgument(admin.isTableAvailable(TableName.valueOf(tableName)), "table " + tableName + " created, but is not available due to some reasons");
        logger.info("create hbase table " + tableName + " done.");
    } finally {
        IOUtils.closeQuietly(admin);
    }

}
 

@Override
public int run(String[] args) throws Exception {
    Options options = new Options();

    options.addOption(OPTION_CUBE_NAME);
    options.addOption(OPTION_PARTITION_FILE_PATH);
    options.addOption(OPTION_HTABLE_NAME);
    parseOptions(options, args);

    Path partitionFilePath = new Path(getOptionValue(OPTION_PARTITION_FILE_PATH));

    String cubeName = getOptionValue(OPTION_CUBE_NAME).toUpperCase();
    KylinConfig config = KylinConfig.getInstanceFromEnv();
    CubeManager cubeMgr = CubeManager.getInstance(config);
    CubeInstance cube = cubeMgr.getCube(cubeName);
    CubeDesc cubeDesc = cube.getDescriptor();

    String tableName = getOptionValue(OPTION_HTABLE_NAME).toUpperCase();
    HTableDescriptor tableDesc = new HTableDescriptor(TableName.valueOf(tableName));
    // https://hbase.apache.org/apidocs/org/apache/hadoop/hbase/regionserver/ConstantSizeRegionSplitPolicy.html
    tableDesc.setValue(HTableDescriptor.SPLIT_POLICY, ConstantSizeRegionSplitPolicy.class.getName());
    tableDesc.setValue(IRealizationConstants.HTableTag, config.getMetadataUrlPrefix());

    Configuration conf = HBaseConfiguration.create(getConf());
    HBaseAdmin admin = new HBaseAdmin(conf);

    try {
        if (User.isHBaseSecurityEnabled(conf)) {
            // add coprocessor for bulk load
            tableDesc.addCoprocessor("org.apache.hadoop.hbase.security.access.SecureBulkLoadEndpoint");
        }

        for (HBaseColumnFamilyDesc cfDesc : cubeDesc.getHBaseMapping().getColumnFamily()) {
            HColumnDescriptor cf = new HColumnDescriptor(cfDesc.getName());
            cf.setMaxVersions(1);

            if (LZOSupportnessChecker.getSupportness()) {
                logger.info("hbase will use lzo to compress data");
                cf.setCompressionType(Algorithm.LZO);
            } else {
                logger.info("hbase will not use lzo to compress data");
            }

            cf.setDataBlockEncoding(DataBlockEncoding.FAST_DIFF);
            cf.setInMemory(false);
            cf.setBlocksize(4 * 1024 * 1024); // set to 4MB
            tableDesc.addFamily(cf);
        }

        byte[][] splitKeys = getSplits(conf, partitionFilePath);

        if (admin.tableExists(tableName)) {
            // admin.disableTable(tableName);
            // admin.deleteTable(tableName);
            throw new RuntimeException("HBase table " + tableName + " exists!");
        }

        DeployCoprocessorCLI.deployCoprocessor(tableDesc);

        admin.createTable(tableDesc, splitKeys);
        logger.info("create hbase table " + tableName + " done.");

        return 0;
    } catch (Exception e) {
        printUsage(options);
        e.printStackTrace(System.err);
        logger.error(e.getLocalizedMessage(), e);
        return 2;
    } finally {
        admin.close();
    }
}
 

@Override
public int run(String[] args) throws Exception {
    Options options = new Options();

    try {
        options.addOption(OPTION_INPUT_PATH);
        options.addOption(OPTION_HTABLE_NAME);
        options.addOption(OPTION_CUBE_NAME);
        parseOptions(options, args);

        String tableName = getOptionValue(OPTION_HTABLE_NAME).toUpperCase();
        // e.g
        // /tmp/kylin-3f150b00-3332-41ca-9d3d-652f67f044d7/test_kylin_cube_with_slr_ready_2_segments/hfile/
        // end with "/"
        String input = getOptionValue(OPTION_INPUT_PATH);

        Configuration conf = HBaseConfiguration.create(getConf());
        FileSystem fs = FileSystem.get(conf);

        String cubeName = getOptionValue(OPTION_CUBE_NAME).toUpperCase();
        KylinConfig config = KylinConfig.getInstanceFromEnv();
        CubeManager cubeMgr = CubeManager.getInstance(config);
        CubeInstance cube = cubeMgr.getCube(cubeName);
        CubeDesc cubeDesc = cube.getDescriptor();
        FsPermission permission = new FsPermission((short) 0777);
        for (HBaseColumnFamilyDesc cf : cubeDesc.getHBaseMapping().getColumnFamily()) {
            String cfName = cf.getName();
            fs.setPermission(new Path(input + cfName), permission);
        }

        String[] newArgs = new String[2];
        newArgs[0] = input;
        newArgs[1] = tableName;

        log.debug("Start to run LoadIncrementalHFiles");
        int ret = ToolRunner.run(new LoadIncrementalHFiles(conf), newArgs);
        log.debug("End to run LoadIncrementalHFiles");
        return ret;
    } catch (Exception e) {
        printUsage(options);
        throw e;
    }
}