Reading space-separated data with spaces data

Question

I'm reading lines from a text file where the data is separated in columns by spaces similar to this:

UNITED STATES OF AMERICA           WASHINGTON          9629047   291289535
CHINA                              PEKING              9596960   1273111290

I had previously handled similar data using the following code:

ifstream readThis("somefile.txt", ios::in);
while (readThis >> country >> capital >> area >> population) {
    // some code...
}

This worked fine when the data didn't have spaces (like "UNITED STATES OF AMERICA"). What happens now is that as soon as a space is encountered the data is saved to the next variable (ie. "2UNITED" would go to country, "STATES" would go to capital and so on). What I'm about to do is what I feel is pretty hack-y so I was hoping they'd be a better way of handling the data. Here's what I think of doing now:

• Read the entire line with std::getline.
• Go through the line character by character.
• Store the characters in the proper variable until we've read 2 spaces in a row.
• At this point ignore any whitespace and read until we reach a character.

This method looks more like an exercise from K&R and probably isn't a C++ way of doing this. I should mention that the data is all properly aligned (the "columns" are all the same width). I'm thinking there has to be a way to read "aligned" data properly (basically the opposite of cout << setw(20) << "Hello" << ...

Any ideas welcomed. Thanks!

Answer 1

I suggest a better method is:
1. read the entire line with std::getline.
2. Extract the fields using std::substr and the field widths.
3. Trim the field strings as necessary.
4. Process the fields. 5. Repeat at step 1 until the read fails.

Answer 2

That is a clear case for regular expressions if I know one (live here):

#include <iostream>
#include <sstream>
#include <boost/regex.hpp>

int main() {
    std::istringstream i { "UNITED STATES OF AMERICA  WASHINGTON, DC  2233232  23232323\nPOPULAR REPUBLIC OF CHINA  BEIJING  23232323  23232323\nBRAZIL  BRASILIA  232323233  2323323\n" };
    boost::regex r { R"(^(.*?)\s\s+(.*?)\s\s+(\d+)\s\s+(\d+))", boost::regex::perl };
    std::string line;
    while( std::getline(i, line) ) {
        boost::smatch m;
        if( !boost::regex_match(line, m, r) )
            continue;
        auto country = m[1].str();
        auto capital = m[2].str();
        auto area = m[3].str();
        auto pop = m[3].str();
        std::cout << capital << ", " << country << ";\n";
    }
}

Notice that

#include <regex>

and the use of std::regex, std::smatch, std::regex_match only work if you are using libc++, the GNU libstdc++ (up to 4.8) isn't working.