Examples with ColumnEncodingKind io.trino.orc.metadata.ColumnEncoding.ColumnEncodingKind used on opensource projects

Example 1 with ColumnEncodingKind

use of io.trino.orc.metadata.ColumnEncoding.ColumnEncodingKind in project trino by trinodb.

the class StripeReader method createValueStreams.

private Map<StreamId, ValueInputStream<?>> createValueStreams(Map<StreamId, Stream> streams, Map<StreamId, OrcChunkLoader> streamsData, ColumnMetadata<ColumnEncoding> columnEncodings) {
    ImmutableMap.Builder<StreamId, ValueInputStream<?>> valueStreams = ImmutableMap.builder();
    for (Entry<StreamId, Stream> entry : streams.entrySet()) {
        StreamId streamId = entry.getKey();
        Stream stream = entry.getValue();
        ColumnEncodingKind columnEncoding = columnEncodings.get(stream.getColumnId()).getColumnEncodingKind();
        // skip index and empty streams
        if (isIndexStream(stream) || stream.getLength() == 0) {
            continue;
        }
        OrcChunkLoader chunkLoader = streamsData.get(streamId);
        OrcTypeKind columnType = types.get(stream.getColumnId()).getOrcTypeKind();
        valueStreams.put(streamId, ValueStreams.createValueStreams(streamId, chunkLoader, columnType, columnEncoding));
    }
    return valueStreams.buildOrThrow();
}

Example 2 with ColumnEncodingKind

use of io.trino.orc.metadata.ColumnEncoding.ColumnEncodingKind in project trino by trinodb.

the class Checkpoints method getStreamCheckpoints.

public static Map<StreamId, StreamCheckpoint> getStreamCheckpoints(Set<OrcColumnId> columns, ColumnMetadata<OrcType> columnTypes, boolean compressed, int rowGroupId, ColumnMetadata<ColumnEncoding> columnEncodings, Map<StreamId, Stream> streams, Map<StreamId, List<RowGroupIndex>> columnIndexes) throws InvalidCheckpointException {
    ImmutableSetMultimap.Builder<OrcColumnId, StreamKind> streamKindsBuilder = ImmutableSetMultimap.builder();
    for (Stream stream : streams.values()) {
        streamKindsBuilder.put(stream.getColumnId(), stream.getStreamKind());
    }
    SetMultimap<OrcColumnId, StreamKind> streamKinds = streamKindsBuilder.build();
    ImmutableMap.Builder<StreamId, StreamCheckpoint> checkpoints = ImmutableMap.builder();
    for (Map.Entry<StreamId, List<RowGroupIndex>> entry : columnIndexes.entrySet()) {
        OrcColumnId columnId = entry.getKey().getColumnId();
        if (!columns.contains(columnId)) {
            continue;
        }
        List<Integer> positionsList = entry.getValue().get(rowGroupId).getPositions();
        ColumnEncodingKind columnEncoding = columnEncodings.get(columnId).getColumnEncodingKind();
        OrcTypeKind columnType = columnTypes.get(columnId).getOrcTypeKind();
        Set<StreamKind> availableStreams = streamKinds.get(columnId);
        ColumnPositionsList columnPositionsList = new ColumnPositionsList(columnId, columnType, positionsList);
        switch(columnType) {
            case BOOLEAN:
                checkpoints.putAll(getBooleanColumnCheckpoints(columnId, compressed, availableStreams, columnPositionsList));
                break;
            case BYTE:
                checkpoints.putAll(getByteColumnCheckpoints(columnId, compressed, availableStreams, columnPositionsList));
                break;
            case SHORT:
            case INT:
            case LONG:
            case DATE:
                checkpoints.putAll(getLongColumnCheckpoints(columnId, columnEncoding, compressed, availableStreams, columnPositionsList));
                break;
            case FLOAT:
                checkpoints.putAll(getFloatColumnCheckpoints(columnId, compressed, availableStreams, columnPositionsList));
                break;
            case DOUBLE:
                checkpoints.putAll(getDoubleColumnCheckpoints(columnId, compressed, availableStreams, columnPositionsList));
                break;
            case TIMESTAMP:
            case TIMESTAMP_INSTANT:
                checkpoints.putAll(getTimestampColumnCheckpoints(columnId, columnEncoding, compressed, availableStreams, columnPositionsList));
                break;
            case BINARY:
            case STRING:
            case VARCHAR:
            case CHAR:
                checkpoints.putAll(getSliceColumnCheckpoints(columnId, columnEncoding, compressed, availableStreams, columnPositionsList));
                break;
            case LIST:
            case MAP:
                checkpoints.putAll(getListOrMapColumnCheckpoints(columnId, columnEncoding, compressed, availableStreams, columnPositionsList));
                break;
            case STRUCT:
                checkpoints.putAll(getStructColumnCheckpoints(columnId, compressed, availableStreams, columnPositionsList));
                break;
            case DECIMAL:
                checkpoints.putAll(getDecimalColumnCheckpoints(columnId, columnEncoding, compressed, availableStreams, columnPositionsList));
                break;
            default:
                throw new IllegalArgumentException("Unsupported column type " + columnType);
        }
    }
    return checkpoints.buildOrThrow();
}

Example 3 with ColumnEncodingKind

use of io.trino.orc.metadata.ColumnEncoding.ColumnEncodingKind in project trino by trinodb.

the class SliceColumnReader method startStripe.

@Override
public void startStripe(ZoneId fileTimeZone, InputStreamSources dictionaryStreamSources, ColumnMetadata<ColumnEncoding> encoding) throws IOException {
    ColumnEncodingKind columnEncodingKind = encoding.get(column.getColumnId()).getColumnEncodingKind();
    if (columnEncodingKind == DIRECT || columnEncodingKind == DIRECT_V2) {
        currentReader = directReader;
    } else if (columnEncodingKind == DICTIONARY || columnEncodingKind == DICTIONARY_V2) {
        currentReader = dictionaryReader;
    } else {
        throw new IllegalArgumentException("Unsupported encoding " + columnEncodingKind);
    }
    currentReader.startStripe(fileTimeZone, dictionaryStreamSources, encoding);
}

Example 4 with ColumnEncodingKind

use of io.trino.orc.metadata.ColumnEncoding.ColumnEncodingKind in project trino by trinodb.

the class StripeReader method createDictionaryStreamSources.

private InputStreamSources createDictionaryStreamSources(Map<StreamId, Stream> streams, Map<StreamId, ValueInputStream<?>> valueStreams, ColumnMetadata<ColumnEncoding> columnEncodings) {
    ImmutableMap.Builder<StreamId, InputStreamSource<?>> dictionaryStreamBuilder = ImmutableMap.builder();
    for (Entry<StreamId, Stream> entry : streams.entrySet()) {
        StreamId streamId = entry.getKey();
        Stream stream = entry.getValue();
        OrcColumnId column = stream.getColumnId();
        // only process dictionary streams
        ColumnEncodingKind columnEncoding = columnEncodings.get(column).getColumnEncodingKind();
        if (!isDictionary(stream, columnEncoding)) {
            continue;
        }
        // skip streams without data
        ValueInputStream<?> valueStream = valueStreams.get(streamId);
        if (valueStream == null) {
            continue;
        }
        OrcTypeKind columnType = types.get(stream.getColumnId()).getOrcTypeKind();
        StreamCheckpoint streamCheckpoint = getDictionaryStreamCheckpoint(streamId, columnType, columnEncoding);
        InputStreamSource<?> streamSource = createCheckpointStreamSource(valueStream, streamCheckpoint);
        dictionaryStreamBuilder.put(streamId, streamSource);
    }
    return new InputStreamSources(dictionaryStreamBuilder.buildOrThrow());
}