在C ++中快速解析制表符分隔的字符串和整数

Question

I have a file which is a couple gigabytes large, and has millions of lines. Each line has data separated like so:

string TAB int TAB int TAB int NEWLINE

My previous attempts to read this line by line have bottle necked as a result of the CPU instead of my SSD's write speed.

How can I quickly parse a massive file line by line?

Note: The files can't be parsed into a vector all at once because they are too large.

In my original code I was parsing the data into vector of structs like this

struct datastruct {
    std::string name;
    int year;
    int occurences;
    int volcount;
};
std::vector<datastruct> data;

Answer 1

Using your datastruct , you could do

std::ifstream file;
datastruct data;
while (file >> data.name >> data.year >> data.occurences >> data.volcount)
{
    // do what you want with data, its contents will be replaced during next iteration
}

Is that that slow?