Examples with BlockletScannedResult org.apache.carbondata.core.scan.result.BlockletScannedResult used on opensource projects

Example 1 with BlockletScannedResult

use of org.apache.carbondata.core.scan.result.BlockletScannedResult in project carbondata by apache.

the class DataBlockIterator method processNextBlocklet.

private BlockletScannedResult processNextBlocklet() throws Exception {
    BlockletScannedResult result = null;
    if (blockExecutionInfo.isPrefetchBlocklet()) {
        if (blockletIterator.hasNext() || nextBlock.get() || nextRead.get()) {
            if (future == null) {
                future = scanNextBlockletAsync();
            }
            result = future.get();
            nextBlock.set(false);
            if (blockletIterator.hasNext() || nextRead.get()) {
                nextBlock.set(true);
                future = scanNextBlockletAsync();
            }
        }
    } else {
        if (blockletIterator.hasNext()) {
            RawBlockletColumnChunks rawChunks = readNextBlockletColumnChunks();
            if (rawChunks != null) {
                result = blockletScanner.scanBlocklet(rawChunks);
            }
        }
    }
    return result;
}

Example 2 with BlockletScannedResult

use of org.apache.carbondata.core.scan.result.BlockletScannedResult in project carbondata by apache.

the class BlockletFilterScanner method executeFilter.

/**
 * This method will process the data in below order
 * 1. first apply min max on the filter tree and check whether any of the filter
 * is fall on the range of min max, if not then return empty result
 * 2. If filter falls on min max range then apply filter on actual
 * data and get the filtered row index
 * 3. if row index is empty then return the empty result
 * 4. if row indexes is not empty then read only those blocks(measure or dimension)
 * which was present in the query but not present in the filter, as while applying filter
 * some of the blocks where already read and present in chunk holder so not need to
 * read those blocks again, this is to avoid reading of same blocks which was already read
 * 5. Set the blocks and filter indexes to result
 *
 * @param rawBlockletColumnChunks
 * @throws FilterUnsupportedException
 */
private BlockletScannedResult executeFilter(RawBlockletColumnChunks rawBlockletColumnChunks) throws FilterUnsupportedException, IOException {
    long startTime = System.currentTimeMillis();
    QueryStatistic totalBlockletStatistic = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.TOTAL_BLOCKLET_NUM);
    totalBlockletStatistic.addCountStatistic(QueryStatisticsConstants.TOTAL_BLOCKLET_NUM, totalBlockletStatistic.getCount() + 1);
    // set the indexed data if it has any during fgdatamap pruning.
    rawBlockletColumnChunks.setBitSetGroup(rawBlockletColumnChunks.getDataBlock().getIndexedData());
    // apply filter on actual data, for each page
    BitSetGroup bitSetGroup = this.filterExecuter.applyFilter(rawBlockletColumnChunks, useBitSetPipeLine);
    // if filter result is empty then return with empty result
    if (bitSetGroup.isEmpty()) {
        CarbonUtil.freeMemory(rawBlockletColumnChunks.getDimensionRawColumnChunks(), rawBlockletColumnChunks.getMeasureRawColumnChunks());
        QueryStatistic scanTime = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.SCAN_BLOCKlET_TIME);
        scanTime.addCountStatistic(QueryStatisticsConstants.SCAN_BLOCKlET_TIME, scanTime.getCount() + (System.currentTimeMillis() - startTime));
        QueryStatistic scannedPages = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.PAGE_SCANNED);
        scannedPages.addCountStatistic(QueryStatisticsConstants.PAGE_SCANNED, scannedPages.getCount() + bitSetGroup.getScannedPages());
        return createEmptyResult();
    }
    BlockletScannedResult scannedResult = new FilterQueryScannedResult(blockExecutionInfo);
    scannedResult.setBlockletId(blockExecutionInfo.getBlockIdString() + CarbonCommonConstants.FILE_SEPARATOR + rawBlockletColumnChunks.getDataBlock().blockletIndex());
    // valid scanned blocklet
    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() + bitSetGroup.getValidPages());
    QueryStatistic scannedPages = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.PAGE_SCANNED);
    scannedPages.addCountStatistic(QueryStatisticsConstants.PAGE_SCANNED, scannedPages.getCount() + bitSetGroup.getScannedPages());
    int[] pageFilteredRowCount = new int[bitSetGroup.getNumberOfPages()];
    // get the row indexes from bit set for each page
    int[][] pageFilteredRowId = new int[bitSetGroup.getNumberOfPages()][];
    int numPages = pageFilteredRowId.length;
    for (int pageId = 0; pageId < numPages; pageId++) {
        BitSet bitSet = bitSetGroup.getBitSet(pageId);
        if (bitSet != null && !bitSet.isEmpty()) {
            int[] matchedRowId = new int[bitSet.cardinality()];
            int index = 0;
            for (int i = bitSet.nextSetBit(0); i >= 0; i = bitSet.nextSetBit(i + 1)) {
                matchedRowId[index++] = i;
            }
            pageFilteredRowCount[pageId] = matchedRowId.length;
            pageFilteredRowId[pageId] = matchedRowId;
        }
    }
    long dimensionReadTime = System.currentTimeMillis();
    dimensionReadTime = System.currentTimeMillis() - dimensionReadTime;
    FileReader fileReader = rawBlockletColumnChunks.getFileReader();
    DimensionRawColumnChunk[] dimensionRawColumnChunks = new DimensionRawColumnChunk[blockExecutionInfo.getTotalNumberDimensionToRead()];
    int numDimensionChunks = dimensionRawColumnChunks.length;
    // read dimension chunk blocks from file which is not present
    for (int chunkIndex = 0; chunkIndex < numDimensionChunks; chunkIndex++) {
        dimensionRawColumnChunks[chunkIndex] = rawBlockletColumnChunks.getDimensionRawColumnChunks()[chunkIndex];
    }
    int[][] allSelectedDimensionColumnIndexRange = blockExecutionInfo.getAllSelectedDimensionColumnIndexRange();
    DimensionRawColumnChunk[] projectionListDimensionChunk = rawBlockletColumnChunks.getDataBlock().readDimensionChunks(fileReader, allSelectedDimensionColumnIndexRange);
    for (int[] columnIndexRange : allSelectedDimensionColumnIndexRange) {
        System.arraycopy(projectionListDimensionChunk, columnIndexRange[0], dimensionRawColumnChunks, columnIndexRange[0], columnIndexRange[1] + 1 - columnIndexRange[0]);
    }
    /*
     * in case projection if the projected dimension are not loaded in the dimensionColumnDataChunk
     * then loading them
     */
    int[] projectionListDimensionIndexes = blockExecutionInfo.getProjectionListDimensionIndexes();
    for (int projectionListDimensionIndex : projectionListDimensionIndexes) {
        if (null == dimensionRawColumnChunks[projectionListDimensionIndex]) {
            dimensionRawColumnChunks[projectionListDimensionIndex] = rawBlockletColumnChunks.getDataBlock().readDimensionChunk(fileReader, projectionListDimensionIndex);
        }
    }
    DimensionColumnPage[][] dimensionColumnPages = new DimensionColumnPage[numDimensionChunks][numPages];
    for (int chunkIndex = 0; chunkIndex < numDimensionChunks; chunkIndex++) {
        if (dimensionRawColumnChunks[chunkIndex] != null) {
            for (int pageId = 0; pageId < numPages; pageId++) {
                dimensionColumnPages[chunkIndex][pageId] = dimensionRawColumnChunks[chunkIndex].decodeColumnPage(pageId);
            }
        }
    }
    MeasureRawColumnChunk[] measureRawColumnChunks = new MeasureRawColumnChunk[blockExecutionInfo.getTotalNumberOfMeasureToRead()];
    int numMeasureChunks = measureRawColumnChunks.length;
    // read the measure chunk blocks which is not present
    for (int chunkIndex = 0; chunkIndex < numMeasureChunks; chunkIndex++) {
        if (null != rawBlockletColumnChunks.getMeasureRawColumnChunks()[chunkIndex]) {
            measureRawColumnChunks[chunkIndex] = rawBlockletColumnChunks.getMeasureRawColumnChunks()[chunkIndex];
        }
    }
    int[][] allSelectedMeasureColumnIndexRange = blockExecutionInfo.getAllSelectedMeasureIndexRange();
    MeasureRawColumnChunk[] projectionListMeasureChunk = rawBlockletColumnChunks.getDataBlock().readMeasureChunks(fileReader, allSelectedMeasureColumnIndexRange);
    for (int[] columnIndexRange : allSelectedMeasureColumnIndexRange) {
        System.arraycopy(projectionListMeasureChunk, columnIndexRange[0], measureRawColumnChunks, columnIndexRange[0], columnIndexRange[1] + 1 - columnIndexRange[0]);
    }
    /*
     * in case projection if the projected measure are not loaded in the ColumnPage
     * then loading them
     */
    int[] projectionListMeasureIndexes = blockExecutionInfo.getProjectionListMeasureIndexes();
    for (int projectionListMeasureIndex : projectionListMeasureIndexes) {
        if (null == measureRawColumnChunks[projectionListMeasureIndex]) {
            measureRawColumnChunks[projectionListMeasureIndex] = rawBlockletColumnChunks.getDataBlock().readMeasureChunk(fileReader, projectionListMeasureIndex);
        }
    }
    ColumnPage[][] measureColumnPages = new ColumnPage[numMeasureChunks][numPages];
    for (int chunkIndex = 0; chunkIndex < numMeasureChunks; chunkIndex++) {
        if (measureRawColumnChunks[chunkIndex] != null) {
            for (int pageId = 0; pageId < numPages; pageId++) {
                measureColumnPages[chunkIndex][pageId] = measureRawColumnChunks[chunkIndex].decodeColumnPage(pageId);
            }
        }
    }
    scannedResult.setDimensionColumnPages(dimensionColumnPages);
    scannedResult.setPageFilteredRowId(pageFilteredRowId);
    scannedResult.setMeasureColumnPages(measureColumnPages);
    scannedResult.setDimRawColumnChunks(dimensionRawColumnChunks);
    scannedResult.setMsrRawColumnChunks(measureRawColumnChunks);
    scannedResult.setPageFilteredRowCount(pageFilteredRowCount);
    // 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 - dimensionReadTime));
    QueryStatistic readTime = queryStatisticsModel.getStatisticsTypeAndObjMap().get(QueryStatisticsConstants.READ_BLOCKlET_TIME);
    readTime.addCountStatistic(QueryStatisticsConstants.READ_BLOCKlET_TIME, readTime.getCount() + dimensionReadTime);
    return scannedResult;
}

Example 3 with BlockletScannedResult

use of org.apache.carbondata.core.scan.result.BlockletScannedResult in project carbondata by apache.

the class BlockletFullScanner method scanBlocklet.

@Override
public BlockletScannedResult scanBlocklet(RawBlockletColumnChunks rawBlockletColumnChunks) throws IOException, FilterUnsupportedException {
    long startTime = System.currentTimeMillis();
    BlockletScannedResult 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() + rawBlockletColumnChunks.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() + rawBlockletColumnChunks.getDataBlock().numberOfPages());
    scannedResult.setBlockletId(blockExecutionInfo.getBlockIdString() + CarbonCommonConstants.FILE_SEPARATOR + rawBlockletColumnChunks.getDataBlock().blockletIndex());
    if (!blockExecutionInfo.isPrefetchBlocklet()) {
        readBlocklet(rawBlockletColumnChunks);
    }
    DimensionRawColumnChunk[] dimensionRawColumnChunks = rawBlockletColumnChunks.getDimensionRawColumnChunks();
    DimensionColumnPage[][] dimensionColumnDataChunks = new DimensionColumnPage[dimensionRawColumnChunks.length][rawBlockletColumnChunks.getDataBlock().numberOfPages()];
    MeasureRawColumnChunk[] measureRawColumnChunks = rawBlockletColumnChunks.getMeasureRawColumnChunks();
    ColumnPage[][] measureColumnPages = new ColumnPage[measureRawColumnChunks.length][rawBlockletColumnChunks.getDataBlock().numberOfPages()];
    scannedResult.setDimensionColumnPages(dimensionColumnDataChunks);
    scannedResult.setMeasureColumnPages(measureColumnPages);
    scannedResult.setDimRawColumnChunks(dimensionRawColumnChunks);
    scannedResult.setMsrRawColumnChunks(measureRawColumnChunks);
    if (blockExecutionInfo.isPrefetchBlocklet()) {
        for (int i = 0; i < dimensionRawColumnChunks.length; i++) {
            if (dimensionRawColumnChunks[i] != null) {
                dimensionColumnDataChunks[i] = dimensionRawColumnChunks[i].decodeAllColumnPages();
            }
        }
        for (int i = 0; i < measureRawColumnChunks.length; i++) {
            if (measureRawColumnChunks[i] != null) {
                measureColumnPages[i] = measureRawColumnChunks[i].decodeAllColumnPages();
            }
        }
    }
    int[] numberOfRows = null;
    if (blockExecutionInfo.getAllSelectedDimensionColumnIndexRange().length > 0) {
        for (int i = 0; i < dimensionRawColumnChunks.length; i++) {
            if (dimensionRawColumnChunks[i] != null) {
                numberOfRows = dimensionRawColumnChunks[i].getRowCount();
                break;
            }
        }
    } else if (blockExecutionInfo.getAllSelectedMeasureIndexRange().length > 0) {
        for (int i = 0; i < measureRawColumnChunks.length; i++) {
            if (measureRawColumnChunks[i] != null) {
                numberOfRows = measureRawColumnChunks[i].getRowCount();
                break;
            }
        }
    }
    // count(*)  case there would not be any dimensions are measures selected.
    if (numberOfRows == null) {
        numberOfRows = new int[rawBlockletColumnChunks.getDataBlock().numberOfPages()];
        for (int i = 0; i < numberOfRows.length; i++) {
            numberOfRows[i] = CarbonV3DataFormatConstants.NUMBER_OF_ROWS_PER_BLOCKLET_COLUMN_PAGE_DEFAULT;
        }
        int lastPageSize = rawBlockletColumnChunks.getDataBlock().numRows() % CarbonV3DataFormatConstants.NUMBER_OF_ROWS_PER_BLOCKLET_COLUMN_PAGE_DEFAULT;
        ;
        if (lastPageSize > 0) {
            numberOfRows[numberOfRows.length - 1] = lastPageSize;
        }
    }
    scannedResult.setPageFilteredRowCount(numberOfRows);
    if (!blockExecutionInfo.isPrefetchBlocklet()) {
        scannedResult.fillDataChunks();
    }
    // 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;
}