Secondary Index

A data structure that makes it eecient to nd a tuple, given the value of its domain (often a key) attribute(s). Hash table is an excellent choice. In practice, a balanced tree structure, called a B-tree, a \multiway BST," is more common. A binary relation with: Domain = values for some particular at-tribute(s) A. Range = pointers to (representations of) tu-ples. (v; p) means that p is a pointer to a tuple whose value in attribute A is v. Supports lookups that specify value of A. Possible implementation of secondary index: hash table with hash function based on Weight, with 17 buckets; h(w) = w mod 17.