I am trying to solve a problem where i have to sort an array based on the frequency of the elements(in case of a tie the one that comes first in the input array should be present first that is i have to maintain the initial order of elements) in O(nlgn) time. I came up with the following logic

• Make an array of objects and initialize its index and value at that index using the given array.
• Now sort the array of objects, using the value.
• Fill their frequencies in the object's frequency field.
• Sort according to frequency and index

My implementation gives an error of error: must use '.*' or '->*' to call pointer-to-member function moreover the error occurs at the 1st instance of stable_sort(). C++ compiler details are g++ 4.9.2

#include<iostream>
#include<algorithm>
using namespace std;

class Element {
    int index, value, frequency;
public:
    Element();
    void setIndex(int);
    void setValue(int);
    void setFrequency(int);
    int getIndex();
    int getValue();
    int getFrequency();
};

class Solution {
public:
    void sortByFrequency(int [], int);
    bool compareByValue(Element *, Element *);
    bool compareByFrequency(Element *, Element *);
};

Element::Element() {
    frequency = index = value = 0;
}

void Element::setIndex(int i) {
    index = i;
}

void Element::setValue(int i) {
    value = i;
}

void Element::setFrequency(int i) {
    frequency = i;
}

int Element::getIndex() {
    return index;
}

int Element::getValue() {
    return value;
}

int Element::getFrequency() {
    return frequency;
}

void Solution::sortByFrequency(int a[], int n) {
    Element obj[n];
    for(register int i = 0; i < n; i++) {
        obj[i].setIndex(i);
        obj[i].setValue(a[i]);
    }
    stable_sort(obj, obj + n, compareByValue);
    obj[0].setFrequency(1);
    for(register int i = 1; i < n; i++) {
        if(obj[i].getValue() == obj[i - 1].getValue()) {
            obj[i].setFrequency(obj[i].getFrequency() + obj[i - 1].getFrequency() + 1);
            obj[i - 1].setFrequency(-1);
            obj[i].setIndex(obj[i - 1].getIndex());
        } else {
            obj[i].setFrequency(1);
        }
    }
    stable_sort(obj, obj + n, compareByFrequency);
    for(register int i = n - 1, j = 0; i >= 0; i--) {
        if(obj[i].getFrequency() != -1) {
            for(register int k = 0; k < obj[i].getFrequency(); k++) {
                a[j++] = obj[i].getValue();
            }
        }
    }
}

bool Solution::compareByValue(Element *a, Element *b) {
    return (a -> getValue() < b -> getValue());
}

bool Solution::compareByFrequency(Element *a, Element *b) {
    if(a -> getFrequency() != b -> getFrequency()) {
        return (a -> getFrequency() < b -> getFrequency());
    } else {
        return (a -> getIndex() > b -> getIndex());
    }
}

int main() {
	int a[] = {2, 5, 2, 6, -1, 9999999, 5, 8, 8, 8};
	int n = sizeof(a)/sizeof(a[0]);
	Solution obj;
	obj.sortByFrequency(a, n);
    for(register int i = 0; i < n; i++) {
        cout << a[i] << " ";
    }
	return 0;
}

I have seen various JAVA programmers use a self defined Input/Output class for faster input/output in JAVA which thy use as a template in their programs and i was wondering if similar utility(having all the features as it's JAVA version) can be made for python users also. I code in Python 2 so i was hoping if anyone can point me in a right direction.