using ifstream, counting elements of each line in a dat file
 

So I have a question regarding ifstream and reading a file.

I have to read in this file, but some of the lines of the file will have missing sections of data. These lines I want to just throw away and then
continue reading the file on the next line until I've read in the whole thing

A small chunk of the data file (which I'm using for testing now...the actual file is like 1 million lines long) looks like so:

Year,Month,Day,Time(hhmmss.mm),Latitude,Longitude,Magnitude,Depth
1973,01,05,053105.80, 33.47, 140.87,4.5, 56
1973,01,05,114837.50, 33.16, 140.91,3.9, 33
1973,01,06,102116.30, 33.49, 140.85,4.2, 33
1973,01,06,112154.70, 33.27, 140.93,4.5, 46
1973,01,06,145552.80, 33.15, 140.71,4.7, 61
1973,01,09,022114.80, 37.81, 141.69,3.5, 59
1973,01,10,075113.60, 33.62, 140.51,4.1,103
1973,01,10,135240.60, 36.84, 137.36,4.4, 33
1973,01,12,074443.90, 33.11, 140.98,3.9, 20
1973,11,30,170127.70, 34.69, 138.57, ,223
1973,01,13,002147.50, 33.28, 141.06,3.9, 40
1973,01,13,231451.10, 33.33, 141.05,3.8, 33
1973,01,15,131515.80, 36.93, 141.93,4.1, 43
1973,01,15,170405.80, 33.37, 140.82,4.9, 58
1973,01,16,181727.80, 35.96, 139.49,4.1, 60
1973,01,16,201134.10, 37.80, 141.71,4.1, 58
1973,01,17,030734.50, 33.28, 140.61,4.0, 58

notice the one line ending with 223 that is missing a longitude value...that line should be thrown away when I read the whole file as it is an incomplete data record. I called the parameter for the file eqFile (it's earthquake data)

So the procedure I use to make this happen is like so:

void foo(string &eqFile)

{

        ifstream inp;

        char thrw = ' ';

        float year,month,day,time,lat,lng,mag,depth;



inp.open(eqFile.c_str(), ifstream::in);

         

    //if the file is not found

    if(inp.fail())

    {

        cout<< "The file \"" << eqFile << "\" does not exist. Exiting... " << endl;

        exit(1);

    }

    //read from the file if it is found

    else

    {

                //strip off the first line. It is unncessary

                do

                {

                        inp.get(thrw);

                        cout << thrw;

                }

                while(thrw != '\n');



                while(!inp.eof())

        {

                        if(inp >> year >> thrw >> month >> thrw >> day >> thrw >> time >>

                                thrw >> lat >> thrw >> lng >> thrw >> mag >> thrw >> depth)

                        {

                                cout << year << "," << month << "," << day << "," << time << ",";

                                cout << lat << "," << lng << "," << mag << "," << depth;

                                cout << endl;

                        }

        }

                

        inp.close(); //close off the file now that it's done being used

    }

}

What this does however is never stops running once it reaches that line where the errored record exists. I need it to skip over the line somehow and continue reading as before

Can anyone tell me what I've got wrong here? There is something I don't get about when ifstream reading encounters an error in a line i suppose

Re: using ifstream, counting elements of each line in a dat file
 

I would use getline() instead of that loop, then parse the line just read

std::strine line;

while( getline(inp,line) )

{



   // check line for correct number of commas

   // count the number of commas -- should be 8 

   // and make sure the last comma is not the last thing on the line

   // if there are not 8 commas or the last comma is the last one on the

   // line, then continue to the top of this loop

   //

   // line is probably ok to use.  So now break it up into individul fields.

   // use std::string's find() to locate the position of the next comma, then    

   // substr() method to extract the text.

   // 

   // or use istringstream() -- but I'm not familar with it so someone else

   // can probably give you good example if you need it.

}

Re: using ifstream, counting elements of each line in a dat file
 

Thanks for the suggestions Ancient Dragon, but...

The above code is nice, but I'm looking for a way to make it work without having to use getline. Getline seems like it is much more work than I should have to do. I feel like a one or two liner could solve my problem, but just don't know enough abotu the dynamics of using inp >> some_var and what happens when it doesn't find some_var. Any tips on making it work the way I have it currently.

Re: using ifstream, counting elements of each line in a dat file
 

There is no quick and easy solution to your problem. As you have already found out just reading the data into integers doesn't solve your problem because the program doesn't distinguish one line from another. Only getline() can do that for you. Once the line is read, the line must be parsed to see if it contains any errors (or missing elements). Again, ifstream's simple extraction operator >> will not do that for you.