Examples with DimensionColumnDataChunk org.apache.carbondata.core.datastore.chunk.DimensionColumnDataChunk used on opensource projects

Example 1 with DimensionColumnDataChunk

use of org.apache.carbondata.core.datastore.chunk.DimensionColumnDataChunk in project carbondata by apache.

the class AbstractBlockletScanner method scanBlocklet.

@Override
public AbstractScannedResult scanBlocklet(BlocksChunkHolder blocksChunkHolder) throws IOException, FilterUnsupportedException {
    long startTime = System.currentTimeMillis();
    AbstractScannedResult scannedResult = new NonFilterQueryScannedResult(blockExecutionInfo);
    QueryStatistic totalBlockletStatistic = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.TOTAL_BLOCKLET_NUM);
    totalBlockletStatistic.addCountStatistic(QueryStatisticsConstants.TOTAL_BLOCKLET_NUM, totalBlockletStatistic.getCount() + 1);
    QueryStatistic validScannedBlockletStatistic = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.VALID_SCAN_BLOCKLET_NUM);
    validScannedBlockletStatistic.addCountStatistic(QueryStatisticsConstants.VALID_SCAN_BLOCKLET_NUM, validScannedBlockletStatistic.getCount() + 1);
    // adding statistics for valid number of pages
    QueryStatistic validPages = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.VALID_PAGE_SCANNED);
    validPages.addCountStatistic(QueryStatisticsConstants.VALID_PAGE_SCANNED, validPages.getCount() + blocksChunkHolder.getDataBlock().numberOfPages());
    // adding statistics for number of pages
    QueryStatistic totalPagesScanned = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.TOTAL_PAGE_SCANNED);
    totalPagesScanned.addCountStatistic(QueryStatisticsConstants.TOTAL_PAGE_SCANNED, totalPagesScanned.getCount() + blocksChunkHolder.getDataBlock().numberOfPages());
    scannedResult.setBlockletId(blockExecutionInfo.getBlockId() + CarbonCommonConstants.FILE_SEPARATOR + blocksChunkHolder.getDataBlock().nodeNumber());
    DimensionRawColumnChunk[] dimensionRawColumnChunks = blocksChunkHolder.getDimensionRawDataChunk();
    DimensionColumnDataChunk[][] dimensionColumnDataChunks = new DimensionColumnDataChunk[dimensionRawColumnChunks.length][];
    for (int i = 0; i < dimensionRawColumnChunks.length; i++) {
        if (dimensionRawColumnChunks[i] != null) {
            dimensionColumnDataChunks[i] = dimensionRawColumnChunks[i].convertToDimColDataChunks();
        }
    }
    scannedResult.setDimensionChunks(dimensionColumnDataChunks);
    MeasureRawColumnChunk[] measureRawColumnChunks = blocksChunkHolder.getMeasureRawDataChunk();
    MeasureColumnDataChunk[][] measureColumnDataChunks = new MeasureColumnDataChunk[measureRawColumnChunks.length][];
    for (int i = 0; i < measureRawColumnChunks.length; i++) {
        if (measureRawColumnChunks[i] != null) {
            measureColumnDataChunks[i] = measureRawColumnChunks[i].convertToMeasureColDataChunks();
        }
    }
    scannedResult.setMeasureChunks(measureColumnDataChunks);
    int[] numberOfRows = new int[] { blocksChunkHolder.getDataBlock().nodeSize() };
    if (blockExecutionInfo.getAllSelectedDimensionBlocksIndexes().length > 0) {
        for (int i = 0; i < dimensionRawColumnChunks.length; i++) {
            if (dimensionRawColumnChunks[i] != null) {
                numberOfRows = dimensionRawColumnChunks[i].getRowCount();
                break;
            }
        }
    } else if (blockExecutionInfo.getAllSelectedMeasureBlocksIndexes().length > 0) {
        for (int i = 0; i < measureRawColumnChunks.length; i++) {
            if (measureRawColumnChunks[i] != null) {
                numberOfRows = measureRawColumnChunks[i].getRowCount();
                break;
            }
        }
    }
    scannedResult.setNumberOfRows(numberOfRows);
    // loading delete data cache in blockexecutioninfo instance
    DeleteDeltaCacheLoaderIntf deleteCacheLoader = new BlockletDeleteDeltaCacheLoader(scannedResult.getBlockletId(), blocksChunkHolder.getDataBlock(), blockExecutionInfo.getAbsoluteTableIdentifier());
    deleteCacheLoader.loadDeleteDeltaFileDataToCache();
    scannedResult.setBlockletDeleteDeltaCache(blocksChunkHolder.getDataBlock().getDeleteDeltaDataCache());
    scannedResult.setRawColumnChunks(dimensionRawColumnChunks);
    // adding statistics for carbon scan time
    QueryStatistic scanTime = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.SCAN_BLOCKlET_TIME);
    scanTime.addCountStatistic(QueryStatisticsConstants.SCAN_BLOCKlET_TIME, scanTime.getCount() + (System.currentTimeMillis() - startTime));
    return scannedResult;
}

Example 2 with DimensionColumnDataChunk

use of org.apache.carbondata.core.datastore.chunk.DimensionColumnDataChunk in project carbondata by apache.

the class PrimitiveQueryType method parseBlocksAndReturnComplexColumnByteArray.

@Override
public void parseBlocksAndReturnComplexColumnByteArray(DimensionRawColumnChunk[] rawColumnChunks, int rowNumber, int pageNumber, DataOutputStream dataOutputStream) throws IOException {
    DimensionColumnDataChunk dataChunk = rawColumnChunks[blockIndex].convertToDimColDataChunk(pageNumber);
    byte[] currentVal = new byte[dataChunk.getColumnValueSize()];
    copyBlockDataChunk(rawColumnChunks, rowNumber, pageNumber, currentVal);
    dataOutputStream.write(currentVal);
}

Example 3 with DimensionColumnDataChunk

use of org.apache.carbondata.core.datastore.chunk.DimensionColumnDataChunk in project carbondata by apache.

the class ExcludeFilterExecuterImpl method applyFilter.

@Override
public BitSetGroup applyFilter(BlocksChunkHolder blockChunkHolder) throws IOException {
    int blockIndex = segmentProperties.getDimensionOrdinalToBlockMapping().get(dimColEvaluatorInfo.getColumnIndex());
    if (null == blockChunkHolder.getDimensionRawDataChunk()[blockIndex]) {
        blockChunkHolder.getDimensionRawDataChunk()[blockIndex] = blockChunkHolder.getDataBlock().getDimensionChunk(blockChunkHolder.getFileReader(), blockIndex);
    }
    DimensionRawColumnChunk dimensionRawColumnChunk = blockChunkHolder.getDimensionRawDataChunk()[blockIndex];
    DimensionColumnDataChunk[] dimensionColumnDataChunks = dimensionRawColumnChunk.convertToDimColDataChunks();
    BitSetGroup bitSetGroup = new BitSetGroup(dimensionRawColumnChunk.getPagesCount());
    for (int i = 0; i < dimensionColumnDataChunks.length; i++) {
        BitSet bitSet = getFilteredIndexes(dimensionColumnDataChunks[i], dimensionRawColumnChunk.getRowCount()[i]);
        bitSetGroup.setBitSet(bitSet, i);
    }
    return bitSetGroup;
}

Example 4 with DimensionColumnDataChunk

use of org.apache.carbondata.core.datastore.chunk.DimensionColumnDataChunk in project carbondata by apache.

the class RowLevelFilterExecuterImpl method readSurrogatesFromColumnBlock.

/**
   * read the filter member dictionary data from the block corresponding to
   * applied filter column
   *
   * @param blockChunkHolder
   * @param index
   * @param dimColumnEvaluatorInfo
   * @return
   */
private int readSurrogatesFromColumnBlock(BlocksChunkHolder blockChunkHolder, int index, int page, DimColumnResolvedFilterInfo dimColumnEvaluatorInfo, int blockIndex) {
    DimensionColumnDataChunk dataChunk = blockChunkHolder.getDimensionRawDataChunk()[blockIndex].convertToDimColDataChunk(page);
    if (dimColumnEvaluatorInfo.getDimension().isColumnar()) {
        byte[] rawData = dataChunk.getChunkData(index);
        ByteBuffer byteBuffer = ByteBuffer.allocate(CarbonCommonConstants.INT_SIZE_IN_BYTE);
        return CarbonUtil.getSurrogateKey(rawData, byteBuffer);
    } else {
        return readSurrogatesFromColumnGroupBlock(dataChunk, index, dimColumnEvaluatorInfo);
    }
}

Example 5 with DimensionColumnDataChunk

use of org.apache.carbondata.core.datastore.chunk.DimensionColumnDataChunk in project carbondata by apache.

the class RowLevelFilterExecuterImpl method createRow.

/**
   * Method will read the members of particular dimension block and create
   * a row instance for further processing of the filters
   *
   * @param blockChunkHolder
   * @param row
   * @param index
   * @throws IOException
   */
private void createRow(BlocksChunkHolder blockChunkHolder, RowIntf row, int pageIndex, int index) throws IOException {
    Object[] record = new Object[dimColEvaluatorInfoList.size() + msrColEvalutorInfoList.size()];
    String memberString;
    for (int i = 0; i < dimColEvaluatorInfoList.size(); i++) {
        DimColumnResolvedFilterInfo dimColumnEvaluatorInfo = dimColEvaluatorInfoList.get(i);
        // if filter dimension is not present in the current add its default value
        if (!isDimensionPresentInCurrentBlock[i]) {
            // fill default value here
            record[dimColumnEvaluatorInfo.getRowIndex()] = getDimensionDefaultValue(dimColumnEvaluatorInfo);
            continue;
        }
        if (dimColumnEvaluatorInfo.getDimension().getDataType() != DataType.ARRAY && dimColumnEvaluatorInfo.getDimension().getDataType() != DataType.STRUCT) {
            if (!dimColumnEvaluatorInfo.isDimensionExistsInCurrentSilce()) {
                record[dimColumnEvaluatorInfo.getRowIndex()] = dimColumnEvaluatorInfo.getDimension().getDefaultValue();
            }
            DimensionColumnDataChunk columnDataChunk = blockChunkHolder.getDimensionRawDataChunk()[dimensionBlocksIndex[i]].convertToDimColDataChunk(pageIndex);
            if (!dimColumnEvaluatorInfo.getDimension().hasEncoding(Encoding.DICTIONARY) && columnDataChunk instanceof VariableLengthDimensionDataChunk) {
                VariableLengthDimensionDataChunk dimensionColumnDataChunk = (VariableLengthDimensionDataChunk) columnDataChunk;
                byte[] memberBytes = dimensionColumnDataChunk.getChunkData(index);
                if (null != memberBytes) {
                    if (Arrays.equals(CarbonCommonConstants.MEMBER_DEFAULT_VAL_ARRAY, memberBytes)) {
                        memberBytes = null;
                    }
                    record[dimColumnEvaluatorInfo.getRowIndex()] = DataTypeUtil.getDataBasedOnDataTypeForNoDictionaryColumn(memberBytes, dimColumnEvaluatorInfo.getDimension().getDataType());
                } else {
                    continue;
                }
            } else {
                int dictionaryValue = readSurrogatesFromColumnBlock(blockChunkHolder, index, pageIndex, dimColumnEvaluatorInfo, dimensionBlocksIndex[i]);
                if (dimColumnEvaluatorInfo.getDimension().hasEncoding(Encoding.DICTIONARY) && !dimColumnEvaluatorInfo.getDimension().hasEncoding(Encoding.DIRECT_DICTIONARY)) {
                    memberString = getFilterActualValueFromDictionaryValue(dimColumnEvaluatorInfo, dictionaryValue);
                    record[dimColumnEvaluatorInfo.getRowIndex()] = DataTypeUtil.getDataBasedOnDataType(memberString, dimColumnEvaluatorInfo.getDimension().getDataType());
                } else if (dimColumnEvaluatorInfo.getDimension().hasEncoding(Encoding.DIRECT_DICTIONARY)) {
                    Object member = getFilterActualValueFromDirectDictionaryValue(dimColumnEvaluatorInfo, dictionaryValue);
                    record[dimColumnEvaluatorInfo.getRowIndex()] = member;
                }
            }
        } else {
            try {
                GenericQueryType complexType = complexDimensionInfoMap.get(dimensionBlocksIndex[i]);
                ByteArrayOutputStream byteStream = new ByteArrayOutputStream();
                DataOutputStream dataOutputStream = new DataOutputStream(byteStream);
                complexType.parseBlocksAndReturnComplexColumnByteArray(blockChunkHolder.getDimensionRawDataChunk(), index, pageIndex, dataOutputStream);
                record[dimColumnEvaluatorInfo.getRowIndex()] = complexType.getDataBasedOnDataTypeFromSurrogates(ByteBuffer.wrap(byteStream.toByteArray()));
                byteStream.close();
            } catch (IOException e) {
                LOGGER.info(e.getMessage());
            }
        }
    }
    DataType msrType;
    for (int i = 0; i < msrColEvalutorInfoList.size(); i++) {
        MeasureColumnResolvedFilterInfo msrColumnEvalutorInfo = msrColEvalutorInfoList.get(i);
        switch(msrColumnEvalutorInfo.getType()) {
            case SHORT:
                msrType = DataType.SHORT;
                break;
            case INT:
                msrType = DataType.INT;
                break;
            case LONG:
                msrType = DataType.LONG;
                break;
            case DECIMAL:
                msrType = DataType.DECIMAL;
                break;
            default:
                msrType = DataType.DOUBLE;
        }
        // in the current block measure list
        if (!isMeasurePresentInCurrentBlock[i]) {
            byte[] defaultValue = msrColumnEvalutorInfo.getCarbonColumn().getDefaultValue();
            record[msrColumnEvalutorInfo.getRowIndex()] = RestructureUtil.getMeasureDefaultValue(msrColumnEvalutorInfo.getCarbonColumn().getColumnSchema(), defaultValue);
            continue;
        }
        Object msrValue;
        MeasureColumnDataChunk measureColumnDataChunk = blockChunkHolder.getMeasureRawDataChunk()[measureBlocksIndex[0]].convertToMeasureColDataChunk(pageIndex);
        switch(msrType) {
            case SHORT:
                msrValue = (short) measureColumnDataChunk.getMeasureDataHolder().getReadableLongValueByIndex(index);
                break;
            case INT:
                msrValue = (int) measureColumnDataChunk.getMeasureDataHolder().getReadableLongValueByIndex(index);
                break;
            case LONG:
                msrValue = measureColumnDataChunk.getMeasureDataHolder().getReadableLongValueByIndex(index);
                break;
            case DECIMAL:
                BigDecimal bigDecimalValue = measureColumnDataChunk.getMeasureDataHolder().getReadableBigDecimalValueByIndex(index);
                if (null != bigDecimalValue && msrColumnEvalutorInfo.getCarbonColumn().getColumnSchema().getScale() > bigDecimalValue.scale()) {
                    bigDecimalValue = bigDecimalValue.setScale(msrColumnEvalutorInfo.getCarbonColumn().getColumnSchema().getScale(), RoundingMode.HALF_UP);
                }
                msrValue = bigDecimalValue;
                break;
            default:
                msrValue = measureColumnDataChunk.getMeasureDataHolder().getReadableDoubleValueByIndex(index);
        }
        record[msrColumnEvalutorInfo.getRowIndex()] = measureColumnDataChunk.getNullValueIndexHolder().getBitSet().get(index) ? null : msrValue;
    }
    row.setValues(record);
}