Collisions

Even if the hash function is randomizing, it is inevitable that sometimes different keys will hash to the same value; this is called a collision.

The load factor λ is the ratio of hash table entries to table size. Obviously, the higher λ is, the greater the probability of a collision.

There are two basic ways to handle a collision:

• Rehash the key using a different hash function. The simplest method of rehashing is linear probing, simply adding 1 to the previous hash value (modulo table size). Other options are quadratic probing and double hashing, using a different hash function. With rehashing, access to the hash table is efficient if the λ is kept below 0.7; this wastes nearly half of the storage.

• Let each table entry point to a bucket of entries, an auxiliary data structure such as a linked list.