Further optimization of finding max in a stack - c++

Question

My problem is to insert an element, pop the last inserted element and printed the max in the given. I am using 2 stacks for the purpose and it is optimized based on most suggested techniques available. But I still need further optimization for the case where number of queries, q <=100000 and test cases<=100. The following is my code as of now:
int main() {
int t,q;
cin>>t;
char query;
int detail;
for(int test=0;test<t;test++)
{
cout<<"Case "<<test+1<<":\n";
cin>>q;
stack<int> s;
stack<int> max;
for(int i=0;i<q;i++)
{
cin>>query;
if(query=='A')
{
cin>>detail;
s.push(detail);
if(max.empty())
max.push(detail);
else if(detail>=max.top())
max.push(detail);
}
else if(query=='R')
{
if(!s.empty())
{
if(s.top()==max.top())
max.pop();
s.pop();
}
}
else
{
if(max.empty())
cout<<"Empty\n";
else
cout<<max.top()<<"\n";
}
}
}
return 0;
}

Answer 1

I'm not sure what do you mean by 'optimize' (minimize memory usage in your cases, improve execution time or just referring to algorithm efficiency). Following comment from #alexeykuzmin0 , the code would look as follows:
std::stack<std::pair<int, int>> stack;
for (int i = 0; i<q; i++)
{
std::cin >> query;
if (query == 'A')
{
std::cin >> detail;
int max = stack.empty() ? detail : std::max(detail, stack.top().second);
stack.push(std::make_pair(detail, max));
}
else if (query == 'R')
{
if (!stack.empty())
{
stack.pop();
}
}
else
{
if (stack.empty())
std::cout << "Empty\n";
else
std::cout << stack.top().second << "\n";
}
}
Additionally it looks like you're not using the history of inserted items for anything, thus we can remove the pair and solve it using single stack.
std::stack<int> stack;
for (int i = 0; i<q; i++)
{
std::cin >> query;
if (query == 'A')
{
std::cin >> detail;
int max = stack.empty() ? detail : std::max(detail, stack.top());
stack.push(max);
}
else if (query == 'R')
{
if (!stack.empty())
{
stack.pop();
}
}
else
{
if (stack.empty())
std::cout << "Empty\n";
else
std::cout << stack.top() << "\n";
}
}
Also remember to measure performance (timings) in release mode builds, with optimizations on. (don't measure performance on debug builds)