An easy-to-use CSV file parser for C++
Andrew DeOrio awdeorio@umich.edu
http://andrewdeorio.com
Table of Contents
- Quick start
- Example 1: Read one column
- Example 2: Read each row and each column with nested loops
- Example 3: Maintaining order of columns in each row
- Changing the delimiter
- Allow too many or too few values in a row
- Error handling
$ git clone https://github.com/awdeorio/csvstream.git
$ cd csvstream/
$ make test
This example reads one column from a CSV file.
// example1.cpp
#include "csvstream.hpp"
#include <iostream>
#include <string>
#include <map>
using namespace std;
int main() {
// Open file
csvstream csvin("input.csv");
// Rows have key = column name, value = cell datum
map<string, string> row;
// Extract the "animal" column
while (csvin >> row) {
cout << row["animal"] << "\n";
}
}
Input
$ cat input.csv
name,animal
Fergie,horse
Myrtle II,chicken
Oscar,cat
Compile
$ make example1
# OR
$ g++ -std=c++11 example1.cpp -o example1
Output
$ ./example1
horse
chicken
cat
This example has an outer loop for each row and an inner loop for each column.
//example2.cpp
#include "csvstream.hpp"
#include <iostream>
#include <string>
#include <map>
using namespace std;
int main() {
// Open file
csvstream csvin("input.csv");
// A row is a map<string, string>, key = column name, value = datum
map<string, string> row;
// Read file
while (csvin >> row) {
cout << "row:" << "\n";
for (auto &col:row) {
const string &column_name = col.first;
const string &datum = col.second;
cout << " " << column_name << ": " << datum << "\n";
}
}
}
Input
$ cat input.csv
name,animal
Fergie,horse
Myrtle II,chicken
Oscar,cat
Compile
$ make example2
# OR
$ g++ -std=c++11 example2.cpp -o example2
Output. Notice output order within each row is animal
followed by name
. This is because iterating over a map
yields items in sorted order by key.
$ ./example2
row:
animal: horse
name: Fergie
row:
animal: chicken
name: Myrtle II
row:
animal: cat
name: Oscar
This example uses a vector-of-pair to maintain the order of values read from each row.
// example3.cpp
#include "csvstream.hpp"
#include <iostream>
#include <string>
#include <vector>
#include <utility>
using namespace std;
int main() {
// Open file
csvstream csvin("input.csv");
// A row is a vector<pair<string, string>>
// key = column name, value = cell datum
vector<pair<string, string>> row;
// Read file
while (csvin >> row) {
cout << "row:" << "\n";
for (unsigned int i=0; i < row.size(); ++i) {
const string &column_name = row[i].first;
const string &datum = row[i].second;
cout << " " << column_name << ": " << datum << "\n";
}
}
}
Input
$ cat input.csv
name,animal
Fergie,horse
Myrtle II,chicken
Oscar,cat
Compile
$ make example3
# OR
$ g++ -std=c++11 example3.cpp -o example3
Output. Notice output order within each row is name
followed by animal
. This is the order that the columns appear in the CSV file.
$ ./example3
row:
name: Fergie
animal: horse
row:
name: Myrtle II
animal: chicken
row:
name: Oscar
animal: cat
By default, values in a row are delimited by a comma ,
. Change the delimiter with the delimiter
constructor parameter.
This example changes the delimiter to the |
character.
csvstream csvin("input.csv", '|');
By default, if a row has too many or too few values, csvstream raises and exception. With strict mode disabled, it will ignore extra values and set missing values to empty string. You must specify a delimiter when using strict mode.
csvstream csvin("input.csv", ',', false);
If an error occurs, csvstream
functions throw a cstream_exception
. For example:
// example4.cpp
#include "csvstream.hpp"
#include <iostream>
#include <string>
#include <map>
using namespace std;
int main() {
// Open file
string filename = "input.csv";
try {
csvstream csvin(filename);
map<string, string> row;
while (csvin >> row) {
cout << row["animal"] << "\n";
}
} catch(const csvstream_exception &e) {
cerr << e.what() << "\n";
return 1;
}
}