Class FactorizedTableIterator

src/include/processor/result/factorized_table.h:270–319

};

class FactorizedTableIterator {
public:
    explicit FactorizedTableIterator(FactorizedTable& factorizedTable);

    bool hasNext() {
        return nextTupleIdx < factorizedTable.getNumTuples() || nextFlatTupleIdx < numFlatTuples;
    }

    void getNext(FlatTuple& tuple);

    void resetState();

private:
    // The dataChunkPos may be not consecutive, which means some entries in the
    // flatTuplePositionsInDataChunk is invalid. We put pair(UINT64_MAX, UINT64_MAX) in the
    // invalid entries.
    bool isValidDataChunkPos(uint32_t dataChunkPos) const {
        return flatTuplePositionsInDataChunk[dataChunkPos].first != UINT64_MAX;
    }

    void readUnflatColToFlatTuple(ft_col_idx_t colIdx, uint8_t* valueBuffer, FlatTuple& tuple);

    void readFlatColToFlatTuple(ft_col_idx_t colIdx, uint8_t* valueBuffer, FlatTuple& tuple);

    // We put pair(UINT64_MAX, UINT64_MAX) in all invalid entries in
    // FlatTuplePositionsInDataChunk.
    void updateInvalidEntriesInFlatTuplePositionsInDataChunk();

    // This function is used to update the number of elements in the dataChunk when we want
    // to iterate a new tuple.
    void updateNumElementsInDataChunk();

    // This function updates the flatTuplePositionsInDataChunk, so that getNextFlatTuple() can
    // correctly outputs the next flat tuple in the current tuple. For example, we want to read
    // two unFlat columns, which are on different dataChunks A,B and both have 100 columns. The
    // flatTuplePositionsInDataChunk after the first call to getNextFlatTuple() looks like:
    // {dataChunkA : [0, 100]}, {dataChunkB : [0, 100]} This function updates the
    // flatTuplePositionsInDataChunk to: {dataChunkA: [1, 100]}, {dataChunkB: [0, 100]}. Meaning
    // that the getNextFlatTuple() should read the second element in the first unflat column and
    // the first element in the second unflat column.
    void updateFlatTuplePositionsInDataChunk();

    const FactorizedTable& factorizedTable;
    uint8_t* currentTupleBuffer;
    uint64_t numFlatTuples;
    ft_tuple_idx_t nextFlatTupleIdx;
    ft_tuple_idx_t nextTupleIdx;
    // This field stores the (nextIdxToReadInDataChunk, numElementsInDataChunk) of each dataChunk.
    std::vector<std::pair<uint64_t, uint64_t>> flatTuplePositionsInDataChunk;
};

} // namespace processor
} // namespace lbug