iterating through elements in unordered map of unordered_maps cpp

Question

Suppose I have an unordered_map defined as below:

unordered_map<int, unordered_map<int, int>> f_table;
f_table[1][3] = 10;
f_table[1][2] = 1;
f_table[1][1] = 2;
f_table[2][3] = 11;
f_table[2][2] = 22;
f_table[2][1] = 4;
f_table[3][3] = 1;
f_table[3][2] = 3;
f_table[3][1] = 2;

And I wanted to sum up all elements in f_table[1] which should add up to 13. How would I go about it?

An unordererd map seems like the wrong tool here. Why not use std::array<std::array<int, 3>, 3> and zero-index? — StoryTeller - Unslander Monica
– StoryTeller - Unslander Monica, Commented Sep 12, 2018 at 7:59
Do you know how to iterate over any other container? Do you know what the value type of a std::unordered_map? When you know that it should become easy. — Some programmer dude
– Some programmer dude, Commented Sep 12, 2018 at 7:59
sum = 0, for(auto& pair : f_table[1]) sum += pair.second; you might want to check if f_table[1] exists first. — Samer Tufail
– Samer Tufail, Commented Sep 12, 2018 at 8:02
How about using vectors then? Or will the "indexes" ever be non-consecutive? — Some programmer dude
– Some programmer dude, Commented Sep 12, 2018 at 8:04

lubgr · Accepted Answer · 2018-09-12 08:47:49Z

7

One approach is using std::accumulate like this:

#include <numeric>

const int result = std::accumulate(f_table[1].cbegin(), f_table[1].cend(),
    0, [](int result, const auto& entry){ return result + entry.second; });

Note that as @StoryTeller pointed out in the comments, you might want to prefer a parallel version of this algorithm, which will ship with fully conforming C++17 implementations, i.e., std::reduce.

Another option is a range based for loop. With structured bindings (again available in C++17), you might consider this more readable:

int result = 0;

for (const auto& [key, value] : f_table[1])
   result += value;

And finally a solution based on range-v3:

#include <range/v3/all.hpp>

using ranges::view::values;
using ranges::accumulate;

const int result = accumulate(f_table[1] | values, 0);

edited Sep 12, 2018 at 8:47

answered Sep 12, 2018 at 8:03

lubgr

38.6k3 gold badges71 silver badges119 bronze badges

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

4 Comments

StoryTeller - Unslander Monica Over a year ago

If we mention C++17, then std::accumulate can be substituted for the paralleled version of std::reduce

lubgr Over a year ago

That's a good idea, but is it available? Doesn't seem to be supported according to this.

StoryTeller - Unslander Monica Over a year ago

Doesn't matter. Answers are not only for the benefit of the OP now, but for future readers too.

Lightness Races in Orbit Over a year ago

Well, it does matter, if the solution doesn't work. There may be some value if it could work in 10 years' time but then post it in 10 years time

JulianW · Accepted Answer · 2018-09-12 08:03:33Z

1

int32_t sum = 0;
for (auto& p : f_table[1])
{
    sum += p.second
}

answered Sep 12, 2018 at 8:03

JulianW

91710 silver badges26 bronze badges

Collectives™ on Stack Overflow

iterating through elements in unordered map of unordered_maps cpp

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related