How to read data from columns not having the same number of rows in HP vertica DB?

Question

I am creating a C++ transform function with the HP Vertica database and I want to read data from 2 columns from 2 different tables and store them into 2 vectors. Then it will output the number of occurences of each element of one vector in the second one. The problem is that these columns do not have the same number of rows. When I logged the code I used I found that it was reading input a much great times than the table size. Here is the code i am using.

  virtual void processPartition(ServerInterface &srvInterface,
                              PartitionReader &inputReader,
                              PartitionWriter &outputWriter)
{

            try {
            std::vector<string> vectcomb;
            std::vector<string> vectrule;
        if (inputReader.getNumCols() != 2)
            vt_report_error(0, "Function only accepts 2 argument, but %zu provided", inputReader.getNumCols());

            try{
            int a=0 ;
        do {


            srvInterface.log("[occurence] data load %d ",a);
            a++;

            if(inputReader.isNull(1)){vectcomb.emplace_back(inputReader.getStringRef(0).str());}
            else {vectcomb.emplace_back(inputReader.getStringRef(0).str());
            vectrule.emplace_back(inputReader.getStringRef(1).str());}

        } while (inputReader.next());
            }catch(exception& e){ srvInterface.log("[occurence] exception catched");}

            for(int i=0;i<vectcomb.size();i++) {
                    int occ=0;
                    srvInterface.log("[occurence] test %d ",i);
                    for(int j=0;j<vectrule.size();j++) {
                            if(vectrule.at(j).find(vectcomb.at(i)) != std::string::npos) {occ++ ;}
                                                              }
                            outputWriter.setInt(0,occ) ; outputWriter.next() ;
                    }

    } catch(exception& e) {
        // Standard exception. Quit.
        vt_report_error(0, "Exception while processing partition: [%s]", e.what());
    }
}
};

score 0 · Answer 1 · answered Aug 04 '16 at 13:24

Something is not make scenes to me , UDX is basically running of top of query result set so I assumed the UDX is getting two columns from the results set as an input , having say that the problem that you refer to is not valid and can’t be exists . It will be very useful if you will be able to provide more info about how you plan to execute this UDX .

How to read data from columns not having the same number of rows in HP vertica DB?

1 Answers1