Notes for week 5 lecture on hash tables

----[ Fri, Mar 14 ]----

Intro to hash tables.

• The USet interface (notes) is supported.
• USet implementation is an important component of transaction systems, such as healthcare.gov. ...a system originally designed for 60,000 simultaneous users but getting 240,000 simultaneous users.
• A hash function is a function from T to hashtable index i in range 0..m-1.
• A hash function allows you to represent the USet as an array called a hash table. The idea is that H.find(x) can return H.t[hash(x)], where H.t is the hash table. ...almost.
• The problem is that two elements x1, x2 may have the same hash value: hash(x1) == hash(x2).
  • Separate Chaining is one way to deal with this "hash clash". ChainedHashTable.h
  • Linear probing is another way to deal with "hash clash". LinearHashTable.h
• We will break hash functions down into two parts:
  Type T object --> unsigned int(range [0..2³²) --> table index(range [0..m)).

  int hash(T x) {
      unsigned int y;
      y = hashCode(x); // from T to range 0..232-1
      return hashFromInt(y); // from 0..232-1 to 0..n-1.
  }
  

  The hashCode is open ended. It depends on the type T and the application.
  The hashFromInt is specific to the implementation of USet interface.