我如何实现MySQL的DISTINCT关键字以使用线程并行获取不同的值

Question

I would like to create a functionality DISTINCT keyword of mysql in c++ using hashmap or unordered_map.

I have to do this in parallel eg initially I have input integers in an array.

Now I have to find distinct number in array (in parallel)

Answer 1

If the original array doesn't change and assuming the available g++ extension for hash_map :

hash_map<int, int> distinct_elems;
for (int i = 0 ; i < num_elems ; ++i)
{
    distinct_elems[i] = i;
}

Since the underlying data doesn't change the distinct_elems won't change either, hence the code:

hash_map<int, int>::iterator de_itr;
for( de_itr = distinct_elems.begin() ; de_itr != distinct_elems.end() ; ++de_itr)
{
    print("%d\n", de_itr->second);
}

or if you simply want to look up the value within the hashmap:

hash_map<int, int>::iterator de_itr = distinct_elems.find(value);
if(de_itr != distinct_elems.end())
{
    <do some work>
}

You can do this from any number of threads with no locking since data is now basically constant.

Answer 2

Use a hashmap and put all the integers there from your array. But add an extra check before you add them to the map so you don't re add them if they exist. At the end, iterate over all the integers in the map and you have your unique items:

Pseudocode:

for (int i=0..array.length){
  if (hashmap.contains(array[i]) continue;
  hashmap.put(i,i);
}

for (iterator over map) {
  print("unique values: " + *it)
}