is it correct to use a tuple as a key for a map data structure in c++?

kikilinux · 05-03-2016, 11:34 PM

I use the bellow code to store IP packets stream in a map data structure.

Code:

typedef std::tuple<uint32, uint32, uint16, uint16, uint8> five_tuple;

std::map<five_tuple, packet> my_map;
// some other code

my_map[five_tuple] = packet;
.
.
.
// ...

auto search = my_map.find (five_tuple variable);

if(search != my_map.end())
     my_map[variable] = new_packet;

Does the above code correct ?

sundialsvcs · 05-05-2016, 08:26 AM

Well, what does it do when you try it?

kikilinux · 05-08-2016, 11:07 PM

It realy work well.
I think the problem of my code is related to this data structure, because I didn't use it before, but it work well.
Another section of my code had bug.

a4z · 05-10-2016, 01:15 AM

Quote:

Originally Posted by kikilinux

It realy work well.
I think the problem of my code is related to this data structure, because I didn't use it before, but it work well.
Another section of my code had bug.

of course it does work well
since std::tuple implements the required operators
http://en.cppreference.com/w/cpp/uti...e/operator_cmp
to fulfil the compare concept,
http://en.cppreference.com/w/cpp/concept/Compare

sundialsvcs · 05-10-2016, 10:33 AM

Incidentally, I normally define a distinct class, in all cases, which descends from Tuple (or whatever). In this way, any logic needed to manipulate or understand the stored items is contained within that class definition itself.