I wonder if there is support in STL for this:
Say I have an class like this :
class Person
{
public:
int getAge() const;
double getIncome() const
Generic adaptor to compare based on member attributes. While it is quite more verbose the first time it is reusable.
// Generic member less than
template <typename T, typename M, typename C>
struct member_lt_type
{
typedef M T::* member_ptr;
member_lt_type( member_ptr p, C c ) : ptr(p), cmp(c) {}
bool operator()( T const & lhs, T const & rhs ) const
{
return cmp( lhs.*ptr, rhs.*ptr );
}
member_ptr ptr;
C cmp;
};
// dereference adaptor
template <typename T, typename C>
struct dereferrer
{
dereferrer( C cmp ) : cmp(cmp) {}
bool operator()( T * lhs, T * rhs ) const {
return cmp( *lhs, *rhs );
}
C cmp;
};
// syntactic sugar
template <typename T, typename M>
member_lt_type<T,M, std::less<M> > member_lt( M T::*ptr ) {
return member_lt_type<T,M, std::less<M> >(ptr, std::less<M>() );
}
template <typename T, typename M, typename C>
member_lt_type<T,M,C> member_lt( M T::*ptr, C cmp ) {
return member_lt_type<T,M,C>( ptr, cmp );
}
template <typename T, typename C>
dereferrer<T,C> deref( C cmp ) {
return dereferrer<T,C>( cmp );
}
// usage:
struct test { int x; }
int main() {
std::vector<test> v;
std::sort( v.begin(), v.end(), member_lt( &test::x ) );
std::sort( v.begin(), v.end(), member_lt( &test::x, std::greater<int>() ) );
std::vector<test*> vp;
std::sort( v.begin(), v.end(), deref<test>( member_lt( &test::x ) ) );
}
This isn't really so much an answer in itself, as a reply to AraK's reply to my comment that sorting with a function instead of a functor can be slower. Here's some (admittedly ugly -- far too much CnP) test code that compares various sorting: qsort, std::sort of vector vs. array, and std::sort using a template class, template function, or plain function for comparison:
#include <vector>
#include <algorithm>
#include <stdlib.h>
#include <time.h>
int compare(void const *a, void const *b) {
if (*(int *)a > *(int *)b)
return -1;
if (*(int *)a == *(int *)b)
return 0;
return 1;
}
const int size = 200000;
typedef unsigned long ul;
void report(char const *title, clock_t ticks) {
printf("%s took %f seconds\n", title, ticks/(double)CLOCKS_PER_SEC);
}
void wait() {
while (clock() == clock())
;
}
template <class T>
struct cmp1 {
bool operator()(T const &a, T const &b) {
return a < b;
}
};
template <class T>
bool cmp2(T const &a, T const &b) {
return a<b;
}
bool cmp3(int a, int b) {
return a<b;
}
int main(void) {
static int array1[size];
static int array2[size];
srand(time(NULL));
for (int i=0; i<size; i++)
array1[i] = rand();
const int iterations = 100;
clock_t total = 0;
for (int i=0; i<iterations; i++) {
memcpy(array2, array1, sizeof(array1));
wait();
clock_t start = clock();
qsort(array2, size, sizeof(array2[0]), compare);
total += clock()-start;
}
report("qsort", total);
total = 0;
for (int i=0; i<iterations; i++) {
memcpy(array2, array1, sizeof(array1));
wait();
clock_t start = clock();
std::sort(array2, array2+size);
total += clock()- start;
}
report("std::sort (array)", total);
total = 0;
for (int i=0; i<iterations; i++) {
memcpy(array2, array1, sizeof(array1));
wait();
clock_t start = clock();
std::sort(array2, array2+size, cmp1<int>());
total += clock()- start;
}
report("std::sort (template class comparator)", total);
total = 0;
for (int i=0; i<iterations; i++) {
memcpy(array2, array1, sizeof(array1));
wait();
clock_t start = clock();
std::sort(array2, array2+size, cmp2<int>);
total += clock()- start;
}
report("std::sort (template func comparator)", total);
total = 0;
for (int i=0; i<iterations; i++) {
memcpy(array2, array1, sizeof(array1));
wait();
clock_t start = clock();
std::sort(array2, array2+size, cmp3);
total += clock()- start;
}
report("std::sort (func comparator)", total);
total = 0;
for (int i=0; i<iterations; i++) {
std::vector<int> array3(array1, array1+size);
wait();
clock_t start = clock();
std::sort(array3.begin(), array3.end());
total += clock()-start;
}
report("std::sort (vector)", total);
return 0;
}
Compiling this with VC++ 9/VS 2008 using cl /O2b2 /GL sortbench3.cpp
, I get:
qsort took 3.393000 seconds
std::sort (array) took 1.724000 seconds
std::sort (template class comparator) took 1.725000 seconds
std::sort (template func comparator) took 2.725000 seconds
std::sort (func comparator) took 2.505000 seconds
std::sort (vector) took 1.721000 seconds
I believe these fall fairly cleanly into three groups: using sort with the default comparison, and using the template class produced the fastest code. Using either the function or template function is clearly slower. Using qsort is (surprisingly to some) the slowest of all, by around a 2:1 margin.
The difference between cmp2 and cmp3 appears to stem entirely from passing by reference vs. value -- if you change cmp2 to take its arguments by value, its speed matches cmp3 exactly (at least in my testing). The difference is that if you know the type is going to be int
, you'll almost certainly pass by value, whereas for generic T
, you'll usually pass by const reference (just in case it's something that's more expensive to copy).
I just tried this based on UncleBens and david-rodriguez-dribeas ideas.
This seems to work (as is) with my current compiler. g++ 3.2.3. Please let me know if it works on other compilers.
#include <vector>
#include <algorithm>
#include <iostream>
using namespace std;
class Person
{
public:
Person( int _age )
:age(_age)
{
}
int getAge() const { return age; }
private:
int age;
};
template <typename T, typename ResType>
class get_lt_type
{
ResType (T::*getter) () const;
public:
get_lt_type(ResType (T::*method) () const ):getter(method) {}
bool operator() ( const T* pT1, const T* pT2 ) const
{
return (pT1->*getter)() < (pT2->*getter)();
}
};
template <typename T, typename ResType>
get_lt_type<T,ResType> get_lt( ResType (T::*getter) () const ) {
return get_lt_type<T, ResType>( getter );
}
int main() {
vector<Person*> people;
people.push_back( new Person( 54 ) );
people.push_back( new Person( 4 ) );
people.push_back( new Person( 14 ) );
sort( people.begin(), people.end(), get_lt( &Person::getAge) );
for ( size_t i = 0; i < people.size(); ++i )
{
cout << people[i]->getAge() << endl;
}
// yes leaking Persons
return 0;
}
If there's only one thing you're ever going to want to sort people by (or if there's a reasonable default that you're going to want to use most of the time), override operator<
for the People
class to sort by this attribute. Without an explicit comparator, STL sorting functions (and anything that makes implicit use of ordering, like sets and maps) will use operator<
.
When you want to sort by something other than operator<
, the way you describe is the only way to do it as of the current version of C++ (although the comparator can just be a regular function; it doesn't have to be a functor). The C++0x standard will make this less verbose by allowing lambda functions.
If you're not willing to wait for C++0x, an alternative is to use boost::lambda.
You don't need to create a class - just write a function:
#include <vector>
#include <algorithm>
using namespace std;
struct Person {
int age;
int getage() const {
return age;
}
};
bool cmp( const Person * a, const Person * b ) {
return a->getage() < b->getage() ;
}
int main() {
vector <Person*> v;
sort( v.begin(), v.end(), cmp );
}
I see that dribeas already posted that idea, but since I already wrote it, here's how you'd write a generic comparator to use getter functions.
#include <functional>
template <class Object, class ResultType>
class CompareAttributeT: public std::binary_function<const Object*, const Object*, bool>
{
typedef ResultType (Object::*Getter)() const;
Getter getter;
public:
CompareAttributeT(Getter method): getter(method) {}
bool operator()(const Object* lhv, const Object* rhv) const
{
return (lhv->*getter)() < (rhv->*getter)();
}
};
template <class Object, class ResultType>
CompareAttributeT<Object, ResultType> CompareAttribute(ResultType (Object::*getter)() const)
{
return CompareAttributeT<Object, ResultType>(getter);
}
Usage:
std::sort(people.begin(), people.end(), CompareAttribute(&Person::getAge));
I think it might be a good idea to overload operator()
for non-pointers, but then one couldn't typedef the argument_types by inheriting from binary_function
- which is probably not a great loss, since it would hard to use it where those are needed anyway, for example, one just couldn't negate the comparison functor anyway.