Just a comparison function? How is, for example, unordered set going to know the end of the array, to hash it? I don't know what is happening under-the-hood so I'm deferring to the list experts. Can you give me a simple example? On Fri, Aug 20, 2010 at 4:31 AM, OvermindDL1 wrote:

On Fri, Aug 20, 2010 at 2:21 AM, B Hart wrote:

...

Can someone please provide an example of how to best use unordered_map/set to store/lookup many small arrays of integers. e.g. unsigned int I[5] = {1000,3344455,12455222,8832232}; I'm guessing that there is no facility for using raw C++ arrays, and that one must wrap them in a struct.

I guess you mean C arrays and not C++ arrays as C++ arrays are probably std::vector's.

C arrays or C++ vectors would work fine though, but you would need to supply your own comparison function, and if C arrays add an ending tag like zero or have a static length. _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

well assuming an unsigned int array size of 5 what do you suggest for a reasonably good and fast hash? I was hoping that I could use the built in hash w/o having to invent one. On Fri, Aug 20, 2010 at 3:32 PM, OvermindDL1 wrote:

Please do not top-post.

On Fri, Aug 20, 2010 at 4:07 PM, B Hart wrote:

...

Just a comparison function? How is, for example, unordered set going to know the end of the array, to hash it? I don't know what is happening under-the-hood so I'm deferring to the list experts.

Can you give me a simple example?

How is it going to know the end? That depends on how you create your array and set up your hash function. If you have an end-of-array marker then you would stop at that (like how '\0' is the end-of-array marker for C strings), if it is a static length then you would just use that, etc... _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

FYI, the implementation of ar_hash doesn't look quite right On Fri, Aug 20, 2010 at 7:52 PM, Tim Moore wrote:

struct ar_hash : std::unary_function {   std::size_t operator()(small_array const& x) const {     int init = 0;     return boost::hash_value(std::accumulate(x.begin(), x.end(), init));   } };

Consider inputs: small_array one = {1,2,3,4}; small_array two = {1,2,2,5}; The sum of each is 10, which will result in the same hash. This will probably cause hash collisions and result in poorer performance than combining each hash value. Replacing the implementation with the following should give better results: return boost::hash_range(x.begin(), x.end()); Nate