不区分大小写的 std::set 字符串

Question

How do you have a case insensitive insertion Or search of a string in std::set?

For example-

std::set<std::string> s;
s.insert("Hello");
s.insert("HELLO"); //not allowed, string already exists.

Answer 1

You need to define a custom comparator:

struct InsensitiveCompare { 
    bool operator() (const std::string& a, const std::string& b) const {
        return strcasecmp(a.c_str(), b.c_str()) < 0;
    }
};

std::set<std::string, InsensitiveCompare> s;

You may try stricmp or strcoll if strcasecmp is not available.

Answer 2

std::set offers the possibility of providing your own comparer (as do most std containers). You can then perform any type of comparison you like. Full example is available here

Answer 3

This is a generic solution that also works with other string types than std::string (tested with std::wstring , std::string_view , char const* ). Basically anything that defines a range of characters should work.

The key point here is to use boost::as_literal that allows us to treat null-terminated character arrays, character pointers and ranges uniformly in the comparator.

Generic code ("iset.h"):

#pragma once
#include <set>
#include <algorithm>
#include <boost/algorithm/string.hpp>
#include <boost/range/as_literal.hpp>

// Case-insensitive generic string comparator.
struct range_iless
{
    template< typename InputRange1, typename InputRange2 >
    bool operator()( InputRange1 const& r1, InputRange2 const& r2 ) const 
    {
        // include the standard begin() and end() aswell as any custom overloads for ADL
        using std::begin; using std::end;  

        // Treat null-terminated character arrays, character pointers and ranges uniformly.
        // This just creates cheap iterator ranges (it doesn't copy container arguments)!
        auto ir1 = boost::as_literal( r1 );
        auto ir2 = boost::as_literal( r2 );

        // Compare case-insensitively.
        return std::lexicographical_compare( 
            begin( ir1 ), end( ir1 ), 
            begin( ir2 ), end( ir2 ), 
            boost::is_iless{} );
    }
};

// Case-insensitive set for any Key that consists of a range of characters.
template< class Key, class Allocator = std::allocator<Key> >
using iset = std::set< Key, range_iless, Allocator >;

Usage example ("main.cpp"):

#include "iset.h"  // above header file
#include <iostream>
#include <string>
#include <string_view>

// Output range to stream.
template< typename InputRange, typename Stream, typename CharT >
void write_to( Stream& s, InputRange const& r, CharT const* sep )
{
    for( auto const& elem : r )
        s << elem << sep;
    s << std::endl;
}

int main()
{
    iset< std::string  >     s1{  "Hello",  "HELLO",  "world" };
    iset< std::wstring >     s2{ L"Hello", L"HELLO", L"world" };
    iset< char const*  >     s3{  "Hello",  "HELLO",  "world" };
    iset< std::string_view > s4{  "Hello",  "HELLO",  "world" };

    write_to( std::cout,  s1,  " " );    
    write_to( std::wcout, s2, L" " );    
    write_to( std::cout,  s3,  " " );    
    write_to( std::cout,  s4,  " " );    
}

Live Demo at Coliru

Answer 4

From what I have read this is more portable than stricmp() because stricmp() is not in fact part of the std library, but only implemented by most compiler vendors. As a result below is my solution to just roll your own.

#include <string>
#include <cctype>
#include <iostream>
#include <set>

struct caseInsensitiveLess
{
  bool operator()(const std::string& x, const std::string& y)
  {
    unsigned int xs ( x.size() );
    unsigned int ys ( y.size() );
    unsigned int bound ( 0 );

    if ( xs < ys ) 
      bound = xs; 
    else 
      bound = ys;

    {
      unsigned int i = 0;
      for (auto it1 = x.begin(), it2 = y.begin(); i < bound; ++i, ++it1, ++it2)
      {
        if (tolower(*it1) < tolower(*it2))
          return true;

        if (tolower(*it2) < tolower(*it1))
          return false;
      }
    }
    return false; 
  }
};

int main()
{
  std::set<std::string, caseInsensitiveLess> ss1;
  std::set<std::string> ss2;

  ss1.insert("This is the first string");
  ss1.insert("THIS IS THE FIRST STRING");
  ss1.insert("THIS IS THE SECOND STRING");
  ss1.insert("This IS THE SECOND STRING");
  ss1.insert("This IS THE Third");

  ss2.insert("this is the first string");
  ss2.insert("this is the first string");
  ss2.insert("this is the second string");
  ss2.insert("this is the second string");
  ss2.insert("this is the third");

  for ( auto& i: ss1 )
   std::cout << i << std::endl;

  std::cout << std::endl;

  for ( auto& i: ss2 )
   std::cout << i << std::endl;

}

Output with case insensitive set and regular set showing the same ordering:

This is the first string
THIS IS THE SECOND STRING
This IS THE Third

this is the first string
this is the second string
this is the third