对C ++ unordered_map和哈希冲突感到困惑

Question

I've read that unordered_map places elements with the same hash in buckets and that's how it handles hash collisions. However, when I checked the insert function , it says:

Each element is inserted only if its key is not equivalent to the key of any other element already in the container

Does that mean that I cannot insert an element with the same hash?.. I'm supposed to be able to insert an element with a new hash since the unordered_map structure can handle collisions, right?.. I think I may be missing something.

Answer 1

It's certainly possible for those statements to be consistent, once you realise that the hash isn't necessarily the key.

A group of distinct keys may generate the same hash value, so be stored in the same bucket, but that still allows for the restriction that duplicate keys are disallowed.

For example, let's say you have a friends collection using first name as the key. The hash function is (a rather simplistic) "use the first letter of the name.

So, while Albert, Andrew, Adam, Bill, Benny and Chloe are six different keys, they only account for three different hash values:

          A                 B            C    (buckets)
   ______/|\_____          / \           |
  /       |      \        /   \          |
Albert  Andrew  Adam    Bill  Benny    Chloe  (keys)