Remove entire rows with missing values c++ - c++

Question

I am reading the data with different variables by the following codes, currently when the program touches missing values (represented in data by string "NA", it will change them to zero. Alternatively, I wonder if how can we remove entire rows when program touch "NA". I have tried to look for the same question but they all are for R, not C++. Please, if you can give me some advises. Thanks
#include <iostream>
#include <fstream>
#include <string>
#include <vector>
using namespace std;
struct Data {
vector<double> cow_id;
vector<double> age_obs;
vector<double> dim_obs;
vector<double> my_obs;
vector<double> mcf_obs;
vector<double> mcp_obs;
vector<double> mcl_obs;
vector<double> bw_obs;
vector<double> bcs_obs;
double get_number (string value)
{
if (value == "NA")
{return 0.0;}
else
{
istringstream iss (value);
double val;
iss>>val;
return val;
}
}
void read_input (const string filepath)
{
ifstream data_in (filepath.c_str());
if (!data_in)
{cout<<"Failed to open"<<endl;}
else
{
// Read tokens as strings.
string id, age, dim, my, mcf, mcp, mcl, bw, bcs;
string dummy_line;
getline(data_in, dummy_line);
string line;
while (data_in >> id >> age >> dim >> my >> mcf >> mcp >> mcl >> bw >> bcs)
{
// Get the number from the string and add to the vectors.
cow_id.push_back(get_number(id));
age_obs.push_back(get_number(age));
dim_obs.push_back(get_number(dim));
my_obs.push_back(get_number(my));
mcf_obs.push_back(get_number(mcf));
mcp_obs.push_back(get_number(mcp));
mcl_obs.push_back(get_number(mcl));
bw_obs.push_back(get_number(bw));
bcs_obs.push_back(get_number(bcs));
}
data_in.close();
}
size_t size=age_obs.size();
for (size_t i=0; i<size; i++)
{
cout<<cow_id[i]<<'\t'<<age_obs[i]<<'\t'<<dim_obs[i]<<'\t'<<my_obs[i] <<'\t'<<mcf_obs[i]<<'\t'<<mcp_obs[i]<<'\t'<<mcl_obs[i]<<'\t'<<bw_obs[i] <<'\t'<<bcs_obs[i]<<endl;
}
};
int main()
{
Data input;
input.read_input("C:\\Data\\C++\\learncpp\\data.txt");
}

Answer 1

Let's talk tables here.
Tables are containers of records (rows). The data you are capturing from your input file is already organized into records. So the obvious model is to use a structure that matches your file's data records.
struct Record
{
unsigned int cow_id;
unsigned int age_obs;
unsigned int dim_obs;
// ...
};
Your table could be represented as:
std::vector<record> my_table;
So to remove a record from the table, you can use the std::vector::erase() method. Easy. Also, you can use the std::find() function to search the table.
Let's relieve some reader's headaches with your present code by introducing a concept of the record loading its members from the file.
Reading a record from a file is best performed by overloading the stream extraction operator>>:
struct Record
{
//...
friend std::istream& operator>>(std::istream& input, Record& r);
};
std::istream&
operator>>(std::istream& input, Record& r)
{
std::string record_text;
std::getline(input, record_text);
// Extract a field from the record text and check for NA,
// Assign fields of r to those values:
r.cow_id = value;
// Etc.
return input;
}
With the overloaded operator, your input looks like:
Record r;
while (input_file >> r)
{
table.push_back(r);
}
Elegant and simple (reducing injection of defects).