from here, i know that we can get the argmax
of a `vector like this:
vector<int> v{ 3, 1, -14, 1, 5, 9 };
vector<int>::iterator result = max_element(v.begin(), v.end());
int argmaxVal = distance(v.begin(), result);
// argmaxVal = 5 because v[5] = 9
How do i get the argmax
of vector from index 2 to 4?
Or, in general, from index k to index n, (assumed to be contiguous).
// argmaxVal = 4 because v[4] = 5
// which is max of v[2], v[3], v[4]
Obviously i can do it like this:
int VectorArgMax(vector<> v, int start_idx, int end_idx){
int retVal = start_idx;
for (int idx=start_idx; idx<=end_idx; idx++){
if v[idx] > v[retVal]{
retVal = idx;
}
}
return retVal;
}
result = VectorArgMax(v,2,4);
But is using max_element
and distance
faster, or any even faster function I can use directly?
You can simply adjust the input iterators to std::max_element
:
std::vector<int>::iterator max = max_element(v.begin() + 2, v.begin() + 4); // [2, 4)
int argmaxVal = distance(v.begin(), max); // absolute index of max
You can also replace std::vector<int>::iterator
with auto
for convenience with C++11
.