unique problem

[s196675m]

HI I am new to this STL form. I have a simple question, unique does not work as expected. Here is my test code

			int iList\[4\] = {0};
			iList\[0\] = -131;
			iList\[1\] = -131;
			iList\[2\] = -343455555;
			iList\[3\] = -131;

			vector<int> iNeighborList(iList, iList + 4);
			vector<int>::iterator it;
			
			it = unique(iNeighborList.begin(), iNeighborList.end());

			iNeighborList.resize(it - iNeighborList.begin());

// At this point
//expected elements of iNeighborList are -131, -343455555
// but actual element I got are -131,-343455555, -131
// Why third element apprear?

			it = remove(iNeighborList.begin(),iNeighborList.end(),-343455555); 
			iNeighborList.resize(it - iNeighborList.begin());

			int isiz = iNeighborList.size();

I am using VS2005 to compile it using Pure MSIL (/clr:pure) switch. Can anybody help me why I am getting uncexppected result. Thank you

[s196675m]

Hi Is there anybody know, what mistake I made in the code . Thank you

[Stuart Dootson]

Your problem is that your expectation of what unique does is incorrect. In the MSDN Documentation for Visual Studio 2008 (and the C++ Standard says the same, but in slightly different language), it says:

Removes duplicate elements that are adjacent to each other in a specified range.

Only the first two elements of your array are of equal value and adjacent to each other.

[s196675m]

Hi Thank you for your good point. Is there any function which can find unique elements in an array vecor even though unique elements may not be adjacent.

[Stuart Dootson]

If the order of elements is unimportant, sort the vector (with std::sort) before using unique. Otherwise, try something like:

#include <set>
#include <vector>
#include <iostream>
#include <iterator>

struct Finder : std::unary_function<bool, int>
{
  Finder(std::set<int>& s) : s_(s) {}
  bool operator()(int i) const
  { 
    if (s_.count(i)==0)
    {
      s_.insert(i);
      return false;
    }
    else
    {
      return true;
    }
  }
  std::set<int>& s_;
};

int main(int, char**)
{
  std::vector<int> data;
  data.push_back(1);
  data.push_back(3);
  data.push_back(2);
  data.push_back(1);
  data.push_back(5);
  data.push_back(1);

  // Show the vector
  std::copy(data.begin(), data.end(), std::ostream_iterator<int>(std::cout, ", ")); std::cout << std::endl;
  
  // Use Finder and a set to remove items if we've seen them before in the vector
  std::set<int> uniqueItems;
  data.erase(std::remove_if(data.begin(), data.end(), Finder(uniqueItems)), data.end());

  // Show the vector with duplicates removed
  std::copy(data.begin(), data.end(), std::ostream_iterator<int>(std::cout, ", ")); std::cout << std::endl;
  return 0;
}

It's not too pretty, but should work OK. And if you parameterise Finder, you can re-use if you need to do it again

[s196675m]

Thanks

[Stephen Hewitt]

The definition of your functor is wrong. You have:

Stuart Dootson wrote:

struct Finder : std::unary_function<bool, int>

It should read:

struct Finder : std::unary_function<int, bool>

Here's the definition of std::unary_function from <functional>:

// TEMPLATE STRUCT unary_function
template<class _A, class _R>
struct unary_function {
typedef _A argument_type;
typedef _R result_type;
};

The result type comes second, not first (which I've always thought was strange). Also, you should be #includeing <functional> for std::unary_function, as things stand you seem to be relying on some other include file including it indirectly. And finally, this technique is not very efficient. Try something like this:

// VanillaConsole.cpp : Defines the entry point for the console application.
//
 
#include "stdafx.h"
#include <iostream>
#include <vector>
#include <algorithm>
#include <iterator>
 
void main()
{
using namespace std;
 
// Create a vector and fill it with data.
typedef vector<int> data_t;
data_t data;
data.push_back(1);
data.push_back(3);
data.push_back(2);
data.push_back(1);
data.push_back(5);
data.push_back(1);
 
// Show the user the contents before messing with it.
cout << "Contents before:" << endl;
copy(data.begin(), data.end(), ostream_iterator<int>(cout, " "));
cout << endl << endl;
 
// Remove the duplicates.
sort(data.begin(), data.end());
data_t::iterator new_end = unique(data.begin(), data.end());
 
// Show the user the contents with duplicates removed (and sorted).
cout < "Contents after:" < endl;
copy(data.begin(), new_end, ostream_iterator<int>(cout, " "));
cout << endl << endl;
}

Steve

[Stephen Hewitt]

Here's[^] how I'd do it.

Steve

[Stuart Dootson]

You might note the first sentence of the post you're replying to:

If the order of elements is unimportant, sort the vector (with std::sort) before using unique.

The whole reason for using the set and stuff is to preserve the order of the elements.

[Stephen Hewitt]

I missed that.

Steve