使用C ++在文件中进行字符串搜索/索引

Question

I am using the following code which searches the file and provides the data and associated line number. But is this code fast enough in case of hundreds of thousands of lines? My PC literally froze for a few seconds. I need to search pair of integers and return its RHS value after comma (some statistical stuff), but with the following code I could able to return the whole line.

1. Is it good idea in terms of fastness to parse the returned data using split functions and get my RHS value

OR

2. Directly get RHS value based on LHS argument. (Well I am unable to do this)

Can anyone help me in achieving any one of the above two?

Here is my code:

#include <string>
#include <iostream>
#include <fstream>

    int main()
    {
        std::ifstream file( "index_hyper.txt" ) ;
        std::string search_str = "401" ;
        std::string line ;
        int line_number = 0 ;
        while( std::getline( file, line ) )
        {
            ++line_number ;
            if( line.find(search_str) != std::string::npos )
                std::cout << "line " << line_number << ": " << line << '\n' ;
        }
    }

Here is my content of index_hyper.txt file:

18,22
20,37
151,61
200,62
156,63
158,64
159,65
153,66
156,67
152,68
154,69
155,56
156,14
157,13
160,122
161,1333
400,455
401,779
402,74
406,71

Answer 1

You can do the work of the code above with:

grep -n "^401," index_hyper.txt

If you want to output just the RHS, you can:

grep  "^401," index_hyper.txt | sed "s/[^,]*,//"

If you are on a Windows platform without sed, grep, bash, etc. then you can easily access unix tools by installing cygwin .

Answer 2

As a general rule, don't start breaking the string up into smaller pieces (substrings) until you need to. And start by specifying exactly what is wanted: you speak of RHS and LHS, and talk of "get RHS value based on LHS argument". So: do you want an exact match on the first field, a substring match on the first field, or a substring match on the entire line?

At any rate: once you have the line in line , you can easily separate it into the two fields:

std::string::const_iterator pivot = std::find( line.cbegin(), line.cend(), ',' );

What you do then depends on what your criterion is:

if ( pivot - line.cbegin() == search_str.size() &&
        std::equal( line.cbegin(), pivot, search_str.begin() ) ) {
    //  Exact match on first field...
    std::cout << std::string( std::next( pivot ), line.cend() );
}

if ( std::search( line.cbegin(), pivot, search_str.begin(), search_str.end() ) != pivot ) {
    //  Matches substring in first field...
    std::cout << std::string( std::next( pivot ), line.cend() );
}

if ( std::search( line.cbegin(), line.cend(), search_str.begin(), search_str.end() ) != line.cend() ) {
    //  Matches substring in complete line...
    std::cout << std::string( std::next( pivot ), line.end() ); }
}

Of course, you'll need some additional error checking. What should you do if there isn't a comma in the line (eg pivot == line.end() ), for example? Or what about extra spaces in the line. (Your example looks like numbers. Should "401" match only "401" , or also "+401" ?)

Before going any further, you should very carefully specify exactly what the code should do, for all possible inputs. (For most possible inputs, of course, the answer will probably be: output an error message with the line number to std::cerr and continue. Being sure to return EXIT_FAILURE in such a case.)