C++ - Remove duplicates from a sorted vector of structs

Question

I have a structure as shown below.

struct Num {
    uint64_t key;
    uint64_t val;
};

// Compare based on their key and value.
bool num_less(Num a, Num b)
{
    if(a.key < b.key)
        return true;
    else if(a.key == b.key)
    {
        if(a.val < b.val)
            return true;
        else
            return false;
    }
    else
        return false;
}

// Compare based on their key and value.
bool num_equal(Num a, Num b)
{
    return ((a.key == b.key) && (a.val == b.val)) ? true : false;
}

I have a vector of structs. I want to remove the duplicates from that vector. I tried the following approach.

Sort the vector
Remove the duplicates(consecutively placed)

vector<Num> arr;

sort(arr.begin(), arr.end(), num_less);

arr.erase(std::unique(arr.begin(), arr.end(), num_less);, arr.end());

But when I run the above code, only the first element of the sorted vector is printed and rest of them are somehow deleted. I am not sure what am I doing wrong.

Edit - I tried with num_equal function in std::unique. And it worked.

You need to provide an overloaded equality operator operator== for your struct. — Dimitrios Bouzas
– Dimitrios Bouzas, Commented Nov 11, 2015 at 13:49
Actually I added the num_equal function in std::unique. But it didn't work. — Pattu
– Pattu, Commented Nov 11, 2015 at 13:51
From en.cppreference.com/w/cpp/algorithm/unique, the 3rd argument is a "binary predicate which returns true if the elements should be treated as equal." So you should use num_equal instead of num_less. — Kevin
– Kevin, Commented Nov 11, 2015 at 13:57
Rule of thumb: only use the constants true and false to initialise variables, and never elsewhere. Your test are terribly convoluted, you can just write return a.val < b.val; instead of if (a.val < b.val) return true; else return false;, for instance. The whole first function should be written as return (a.key < b.key) or (a.key == b.key and a.val < b.val); Equivalently, the second would be return (a.key == b.key) and (a.val == b.val); — Konrad Rudolph
– Konrad Rudolph, Commented Nov 11, 2015 at 14:07

Bushuev · Accepted Answer · 2015-11-12 14:32:28Z

2

you need to define equal predicate

struct Num 
{
    unsigned int key;
    unsigned int val;
};

bool num_less(const Num &a, const Num &b)
{
    return (a.key<b.key)||(!(b.key < a.key))&&(a.val<b.val));
}
bool num_equal(const Num &a, const Num &b)
{
    return (a.key==b.key)&&(a.val==b.val);
}
int main()
{
    vector<Num> arr;
    Num temp;
    // add some examples
    temp.key=10; temp.val=20;
    arr.push_back(temp);        arr.push_back(temp);
    temp.key=11; temp.val=23;
    arr.push_back(temp);        arr.push_back(temp);
    temp.key=10; temp.val=20;
    arr.push_back(temp);        arr.push_back(temp);    arr.push_back(temp);
    //sort
    sort(arr.begin(),arr.end(),num_less);
    //delete dublicates
    arr.erase(unique(arr.begin(),arr.end(),num_equal),arr.end());
    return 0;
}

edited Nov 12, 2015 at 14:32

answered Nov 11, 2015 at 14:05

Bushuev

5771 gold badge12 silver badges31 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

Jarod42 Over a year ago

Your num_less is wrong and can be rewritten correctly as return std::tie(a.key, a.val) < std::tie(b.key, b.val);

Bushuev Over a year ago

you are right it is num_great, but the code works correctly, because of it deletes all dublicats.

Pattu Over a year ago

Yeah, my num_less was messed up. I changed it and it's working fine now.

Jarod42 Over a year ago

@h.o.m.a.n: currently, your num_less is equivalent to (a.key < b.key) ||(a.val < b.val), so {1, 2} < {2, 1} but also {2, 1} < {1, 2}...

Jarod42 Over a year ago

So you have a typo in !(a.key < b.key) which should be !(b.key < a.key).

|

Collectives™ on Stack Overflow

C++ - Remove duplicates from a sorted vector of structs

1 Answer 1

6 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

6 Comments

Your Answer

Sign up or log in

Post as a guest

Related