How does one correctly store data into an array struct with stringstream? [duplicate] - c++

Question

This question already has answers here:
Why does reading a record struct fields from std::istream fail, and how can I fix it?
(9 answers)
Closed 6 years ago.
I was wondering how to store data from a CSV file into a structured array. I realize I need to use getline and such and so far I have come up with this code:
This is my struct:
struct csvData //creating a structure
{
string username; //creating a vector of strings called username
float gpa; //creating a vector of floats called gpa
int age; //creating a vector of ints called age
};
This is my data reader and the part that stores the data:
csvData arrayData[10];
string data;
ifstream infile; //creating object with ifstream
infile.open("datafile.csv"); //opening file
if (infile.is_open()) //error check
int i=0;
while(getline(infile, data));
{
stringstream ss(data);
ss >> arrayData[i].username;
ss >> arrayData[i].gpa;
ss >> arrayData[i].age;
i++;
}
Further, this is how I was attempting to print out the information:
for (int z = 0; z<10; z++)
{
cout<<arrayData[z].username<<arrayData[z].gpa<<arrayData[z].age<<endl;
}
However, when running this command, I get a cout of what seem to be random numbers:
1.83751e-0383 03 4.2039e-0453 1.8368e-0383 07011688
I assume this has to be the array running not storing the variables correctly and thus I am reading out random memory slots, however, I am unsure.
Lastly, here is the CSV file I am attempting to read.
username,gpa,age
Steven,3.2,20
Will,3.4,19
Ryan,3.6,19
Tom,3,19

Answer 1

There's nothing in your parsing code that actually attempts to parse the single line into the individual fields:
while(getline(infile, data));
{
This correctly reads a single line from the input file into the data string.
stringstream ss(data);
ss >> arrayData[i].username;
ss >> arrayData[i].gpa;
ss >> arrayData[i].age;
You need to try to explain to your rubber duck how this is supposed to take a single line of comma-separated values, like the one you showed in your question:
Steven,3.2,20
and separate that string into the individual values, by commas. There's nothing about the >> operator that will do this. operator>> separates input using whitespaces, not commas. Your suspicions were correct, you were not parsing the input correctly.
This is a task that you have to do yourself. I am presuming that you would like, as a learning experience, or as a homework assignment, to do this yourself, manually. Well, then, do it yourself. You have the a single line in data. Use any number of tools that C++ gives you: the std::string's find() method, or std::find() from <algorithm>, to find each comma in the data string, then extract each individual portion of the string that's between each comma. Then, you still need to convert the two numeric fields into the appropriate datatypes. And that's when you put each one of them into a std::istringstream, and use operator>> to convert them to numeric types.
But, having said all that, there's an alternative dirty trick, to solve this problem quickly. Recall that the original line in data contains
Steven,3.2,20
All you have to do is replace the commas with spaces, turning it into:
Steven 3.2 20
Replacing commas with spaces is trivial with std::replace(), or with a small loop. Then, you can stuff the result into a std::istringstream, and use operator>> to extract the individual whitespace-delimited values into the discrete variables, using the code that you've already written.
Just a small word of warning: if this was indeed your homework assignment, to write code to manually parse and extract comma-delimited values, it's not guaranteed that your instructor will give you the full grade for taking the dirty-trick approach...

Answer 2

UNDER CONSTRUCTION
Ton, nice try and nice complete question. Here is the answer:
1) You have a semicolon after the loop:
while(getline(infile, data));
delete it.
How did I figure that out easily? I compiled with all the warnings enabled, like this:
C02QT2UBFVH6-lm:~ gsamaras$ g++ -Wall main.cpp
main.cpp:24:33: warning: while loop has empty body [-Wempty-body]
while(getline(infile, data));
^
main.cpp:24:33: note: put the semicolon on a separate line to silence this warning
1 warning generated.
In fact, you should get that warning without -Wall as well, but get into using it, it will also make good to you! :)
2) Then, you read some elements, but not 10, so why do you print 10? Print as many as the ones you actually read, i.e. i.
When you try to print all 10 elements of your array, you print elements that are not initialized, since you didn't initialize your array of structs.
Moreover, the number of lines in datafile.csv was less than 10. So you started populating your array, but you stopped, when the file didn't have more lines. As a result, some of the elements of your array (the last 6 elements) remained uninitialized.
Printing uninitialized data, causes Undefined Behavior, that's why you see garbage values.
3) Also this:
if (infile.is_open()) //error check
could be written like this:
if (!infile.is_open())
cerr << "Error Message by Mr. Tom\n";
Putting them all together:
WILL STILL NOT WORK, BECAUSE ss >> arrayData[i].username; eats the entire input line and the next two extractions fail, as Pete Becker said, but I leave it here, so that others won't make the same attempt!!!!!!!
#include <iostream>
#include <fstream>
#include <string>
#include <sstream>
using namespace std;
struct csvData //creating a structure
{
string username; //creating a vector of strings called username
float gpa; //creating a vector of floats called gpa
int age; //creating a vector of ints called age
};
int main()
{
csvData arrayData[10];
string data;
ifstream infile; //creating object with ifstream
infile.open("datafile.csv"); //opening file
if (!infile.is_open()) { cerr << "File is not opened..\n"; }
int i=0;
while(getline(infile, data))
{
stringstream ss(data);
ss >> arrayData[i].username;
ss >> arrayData[i].gpa;
ss >> arrayData[i].age;
i++;
}
for (int z = 0; z< i; z++)
{
cout<<arrayData[z].username<<arrayData[z].gpa<<arrayData[z].age<<endl;
}
return 0;
}
Output:
C02QT2UBFVH6-lm:~ gsamaras$ g++ -Wall main.cpp
C02QT2UBFVH6-lm:~ gsamaras$ ./a.out
username,gpa,age00
Steven,3.2,2000
Will,3.4,1900
Ryan,3.6,1900
Tom,3,1900
But wait a minute, so now it works, but why this:
while(getline(infile, data));
{
...
}
didn't?
Because, putting a semicolon after a loop is equivalent to this:
while()
{
;
}
because as you probably already know loops with only one line as a body do not require curly brackets.
And what happened to what I thought it was the body of the loop (i.e. the part were you use std::stringstream)?
It got executed! But only once!.
You see, a pair of curly brackets alone means something, it's an anonymous scope/block.
So this:
{
stringstream ss(data);
ss >> arrayData[i].username;
ss >> arrayData[i].gpa;
ss >> arrayData[i].age;
i++;
}
functioned on its one, without being part of the while loop, as you intended too!
Any why did it work?! Because you had declared i before the loop! ;)