Examples with ScanBatch org.apache.drill.exec.physical.impl.ScanBatch used on opensource projects

Example 16 with ScanBatch

use of org.apache.drill.exec.physical.impl.ScanBatch in project drill by apache.

the class MapRDBRestrictedScanBatchCreator method getBatch.

@Override
public ScanBatch getBatch(ExecutorFragmentContext context, RestrictedMapRDBSubScan subScan, List<RecordBatch> children) throws ExecutionSetupException {
    Preconditions.checkArgument(children.isEmpty());
    List<RecordReader> readers = Lists.newArrayList();
    for (MapRDBSubScanSpec scanSpec : subScan.getRegionScanSpecList()) {
        try {
            readers.add(new RestrictedJsonRecordReader(scanSpec, subScan.getFormatPlugin(), subScan.getColumns(), context, subScan.getMaxRecordsToRead(), subScan.getSchema()));
        } catch (Exception e1) {
            throw new ExecutionSetupException(e1);
        }
    }
    return new ScanBatch(subScan, context, readers, true);
}

Example 17 with ScanBatch

use of org.apache.drill.exec.physical.impl.ScanBatch in project drill by apache.

the class HBaseScanBatchCreator method getBatch.

@Override
public ScanBatch getBatch(ExecutorFragmentContext context, HBaseSubScan subScan, List<RecordBatch> children) throws ExecutionSetupException {
    Preconditions.checkArgument(children.isEmpty());
    List<RecordReader> readers = new LinkedList<>();
    List<SchemaPath> columns = null;
    for (HBaseSubScan.HBaseSubScanSpec scanSpec : subScan.getRegionScanSpecList()) {
        try {
            if ((columns = subScan.getColumns()) == null) {
                columns = GroupScan.ALL_COLUMNS;
            }
            readers.add(new HBaseRecordReader(subScan.getStorageEngine().getConnection(), scanSpec, columns));
        } catch (Exception e1) {
            throw new ExecutionSetupException(e1);
        }
    }
    return new ScanBatch(subScan, context, readers);
}

Example 18 with ScanBatch

use of org.apache.drill.exec.physical.impl.ScanBatch in project drill by apache.

the class HBaseScanBatchCreator method getBatch.

@Override
public ScanBatch getBatch(FragmentContext context, HBaseSubScan subScan, List<RecordBatch> children) throws ExecutionSetupException {
    Preconditions.checkArgument(children.isEmpty());
    List<RecordReader> readers = Lists.newArrayList();
    List<SchemaPath> columns = null;
    for (HBaseSubScan.HBaseSubScanSpec scanSpec : subScan.getRegionScanSpecList()) {
        try {
            if ((columns = subScan.getColumns()) == null) {
                columns = GroupScan.ALL_COLUMNS;
            }
            readers.add(new HBaseRecordReader(subScan.getStorageEngine().getConnection(), scanSpec, columns, context));
        } catch (Exception e1) {
            throw new ExecutionSetupException(e1);
        }
    }
    return new ScanBatch(subScan, context, readers.iterator());
}

Example 19 with ScanBatch

use of org.apache.drill.exec.physical.impl.ScanBatch in project drill by apache.

the class EasyFormatPlugin method getReaderBatch.

@SuppressWarnings("resource")
CloseableRecordBatch getReaderBatch(FragmentContext context, EasySubScan scan) throws ExecutionSetupException {
    final ImplicitColumnExplorer columnExplorer = new ImplicitColumnExplorer(context, scan.getColumns());
    if (!columnExplorer.isStarQuery()) {
        scan = new EasySubScan(scan.getUserName(), scan.getWorkUnits(), scan.getFormatPlugin(), columnExplorer.getTableColumns(), scan.getSelectionRoot());
        scan.setOperatorId(scan.getOperatorId());
    }
    OperatorContext oContext = context.newOperatorContext(scan);
    final DrillFileSystem dfs;
    try {
        dfs = oContext.newFileSystem(fsConf);
    } catch (IOException e) {
        throw new ExecutionSetupException(String.format("Failed to create FileSystem: %s", e.getMessage()), e);
    }
    List<RecordReader> readers = Lists.newArrayList();
    List<Map<String, String>> implicitColumns = Lists.newArrayList();
    Map<String, String> mapWithMaxColumns = Maps.newLinkedHashMap();
    for (FileWork work : scan.getWorkUnits()) {
        RecordReader recordReader = getRecordReader(context, dfs, work, scan.getColumns(), scan.getUserName());
        readers.add(recordReader);
        Map<String, String> implicitValues = columnExplorer.populateImplicitColumns(work, scan.getSelectionRoot());
        implicitColumns.add(implicitValues);
        if (implicitValues.size() > mapWithMaxColumns.size()) {
            mapWithMaxColumns = implicitValues;
        }
    }
    // all readers should have the same number of implicit columns, add missing ones with value null
    Map<String, String> diff = Maps.transformValues(mapWithMaxColumns, Functions.constant((String) null));
    for (Map<String, String> map : implicitColumns) {
        map.putAll(Maps.difference(map, diff).entriesOnlyOnRight());
    }
    return new ScanBatch(scan, context, oContext, readers.iterator(), implicitColumns);
}

Example 20 with ScanBatch

use of org.apache.drill.exec.physical.impl.ScanBatch in project drill by apache.

the class KuduScanBatchCreator method getBatch.

@Override
public ScanBatch getBatch(FragmentContext context, KuduSubScan subScan, List<RecordBatch> children) throws ExecutionSetupException {
    Preconditions.checkArgument(children.isEmpty());
    List<RecordReader> readers = Lists.newArrayList();
    List<SchemaPath> columns = null;
    for (KuduSubScan.KuduSubScanSpec scanSpec : subScan.getTabletScanSpecList()) {
        try {
            if ((columns = subScan.getColumns()) == null) {
                columns = GroupScan.ALL_COLUMNS;
            }
            readers.add(new KuduRecordReader(subScan.getStorageEngine().getClient(), scanSpec, columns, context));
        } catch (Exception e1) {
            throw new ExecutionSetupException(e1);
        }
    }
    return new ScanBatch(subScan, context, readers.iterator());
}