·

1.095K

· kybernetyk

From time to time you'd like to perform an oldschool lispy map on a std::vector and get a new vector as a result. The problem is that std::transform will mutate your original vector if you don't create an explicit output vector and pass its iterator to std::transform.

IMHO that's a little too much work for every time you want to do a simple map() and so I'v built a little wrapper ...

#include <algorithm>

template <typename T, typename OP>
T map(const T& coll, OP lambda) {
 T new_coll(coll.size());

 std::transform(std::begin(coll),
 std::end(coll),
 std::begin(new_coll),
 lambda);

 return new_coll;
}

void testmap() {
 std::vector<Float32> vec = {1.0, 2.0, 3.0};

 //we could also use a [](const Type&) lambda here
 auto vec2 = map(vec, [](Float32 f) -> Float32 {
 return f * 2.0;
 });

 printf("orig data:\t{");
 for (auto f : vec) {
 printf("%f, ", f);
 }
 printf("}\n new data:\t{"); 
 for (auto f : vec2) {
 printf("%f, ", f);
 }
 printf("}\n");
}

Output will be:

orig data: {1.000000, 2.000000, 3.000000, }
 new data: {2.000000, 4.000000, 6.000000, }

Have fun ;)

Related protips

Have a fresh tip? Share with Coderwall community!

Post

Post a tip

Sponsored by

#native_company#

#native_desc#

#native_cta#

Post a job for only 299ドル

Sponsored by #native_company# — Learn More