Search in sources :

Example 6 with FilterResolverIntf

use of org.apache.carbondata.core.scan.filter.resolver.FilterResolverIntf in project carbondata by apache.

the class CarbonTableReader method getInputSplits2.

public List<CarbonLocalInputSplit> getInputSplits2(CarbonTableCacheModel tableCacheModel, Expression filters) throws Exception {
    // need apply filters to segment
    FilterExpressionProcessor filterExpressionProcessor = new FilterExpressionProcessor();
    AbsoluteTableIdentifier absoluteTableIdentifier = tableCacheModel.carbonTable.getAbsoluteTableIdentifier();
    CacheClient cacheClient = new CacheClient(absoluteTableIdentifier.getStorePath());
    List<String> invalidSegments = new ArrayList<>();
    List<UpdateVO> invalidTimestampsList = new ArrayList<>();
    // get all valid segments and set them into the configuration
    SegmentUpdateStatusManager updateStatusManager = new SegmentUpdateStatusManager(absoluteTableIdentifier);
    SegmentStatusManager segmentStatusManager = new SegmentStatusManager(absoluteTableIdentifier);
    SegmentStatusManager.ValidAndInvalidSegmentsInfo segments = segmentStatusManager.getValidAndInvalidSegments();
    tableCacheModel.segments = segments.getValidSegments().toArray(new String[0]);
    if (segments.getValidSegments().size() == 0) {
        return new ArrayList<>(0);
    }
    // remove entry in the segment index if there are invalid segments
    invalidSegments.addAll(segments.getInvalidSegments());
    for (String invalidSegmentId : invalidSegments) {
        invalidTimestampsList.add(updateStatusManager.getInvalidTimestampRange(invalidSegmentId));
    }
    if (invalidSegments.size() > 0) {
        List<TableSegmentUniqueIdentifier> invalidSegmentsIds = new ArrayList<>(invalidSegments.size());
        for (String segId : invalidSegments) {
            invalidSegmentsIds.add(new TableSegmentUniqueIdentifier(absoluteTableIdentifier, segId));
        }
        cacheClient.getSegmentAccessClient().invalidateAll(invalidSegmentsIds);
    }
    // get filter for segment
    CarbonInputFormatUtil.processFilterExpression(filters, tableCacheModel.carbonTable);
    FilterResolverIntf filterInterface = CarbonInputFormatUtil.resolveFilter(filters, tableCacheModel.carbonTable.getAbsoluteTableIdentifier());
    List<CarbonLocalInputSplit> result = new ArrayList<>();
    // for each segment fetch blocks matching filter in Driver BTree
    for (String segmentNo : tableCacheModel.segments) {
        try {
            List<DataRefNode> dataRefNodes = getDataBlocksOfSegment(filterExpressionProcessor, absoluteTableIdentifier, tableCacheModel.carbonTablePath, filterInterface, segmentNo, cacheClient, updateStatusManager);
            for (DataRefNode dataRefNode : dataRefNodes) {
                BlockBTreeLeafNode leafNode = (BlockBTreeLeafNode) dataRefNode;
                TableBlockInfo tableBlockInfo = leafNode.getTableBlockInfo();
                if (CarbonUtil.isInvalidTableBlock(tableBlockInfo, updateStatusManager.getInvalidTimestampRange(tableBlockInfo.getSegmentId()), updateStatusManager)) {
                    continue;
                }
                result.add(new CarbonLocalInputSplit(segmentNo, tableBlockInfo.getFilePath(), tableBlockInfo.getBlockOffset(), tableBlockInfo.getBlockLength(), Arrays.asList(tableBlockInfo.getLocations()), tableBlockInfo.getBlockletInfos().getNoOfBlockLets(), tableBlockInfo.getVersion().number()));
            }
        } catch (Exception ex) {
            throw new RuntimeException(ex);
        }
    }
    cacheClient.close();
    return result;
}
Also used : CacheClient(org.apache.carbondata.hadoop.CacheClient) SegmentUpdateStatusManager(org.apache.carbondata.core.statusmanager.SegmentUpdateStatusManager) SegmentStatusManager(org.apache.carbondata.core.statusmanager.SegmentStatusManager) UpdateVO(org.apache.carbondata.core.mutate.UpdateVO) KeyGenException(org.apache.carbondata.core.keygenerator.KeyGenException) IndexBuilderException(org.apache.carbondata.core.datastore.exception.IndexBuilderException) IOException(java.io.IOException) FilterExpressionProcessor(org.apache.carbondata.core.scan.filter.FilterExpressionProcessor) AbsoluteTableIdentifier(org.apache.carbondata.core.metadata.AbsoluteTableIdentifier) FilterResolverIntf(org.apache.carbondata.core.scan.filter.resolver.FilterResolverIntf) BlockBTreeLeafNode(org.apache.carbondata.core.datastore.impl.btree.BlockBTreeLeafNode)

Aggregations

FilterResolverIntf (org.apache.carbondata.core.scan.filter.resolver.FilterResolverIntf)6 AbsoluteTableIdentifier (org.apache.carbondata.core.metadata.AbsoluteTableIdentifier)4 Expression (org.apache.carbondata.core.scan.expression.Expression)4 CarbonTable (org.apache.carbondata.core.metadata.schema.table.CarbonTable)3 UpdateVO (org.apache.carbondata.core.mutate.UpdateVO)3 IOException (java.io.IOException)2 FilterExpressionProcessor (org.apache.carbondata.core.scan.filter.FilterExpressionProcessor)2 CarbonQueryPlan (org.apache.carbondata.core.scan.model.CarbonQueryPlan)2 QueryModel (org.apache.carbondata.core.scan.model.QueryModel)2 SegmentStatusManager (org.apache.carbondata.core.statusmanager.SegmentStatusManager)2 SegmentUpdateStatusManager (org.apache.carbondata.core.statusmanager.SegmentUpdateStatusManager)2 InputSplit (org.apache.hadoop.mapreduce.InputSplit)2 LinkedList (java.util.LinkedList)1 TableSegmentUniqueIdentifier (org.apache.carbondata.core.datastore.TableSegmentUniqueIdentifier)1 IndexBuilderException (org.apache.carbondata.core.datastore.exception.IndexBuilderException)1 BlockBTreeLeafNode (org.apache.carbondata.core.datastore.impl.btree.BlockBTreeLeafNode)1 KeyGenException (org.apache.carbondata.core.keygenerator.KeyGenException)1 PartitionInfo (org.apache.carbondata.core.metadata.schema.PartitionInfo)1 Partitioner (org.apache.carbondata.core.scan.partition.Partitioner)1 CacheClient (org.apache.carbondata.hadoop.CacheClient)1