DATA3404 Lecture Notes - Lecture 6: Relational Algebra, Memory Stick, Multi-Core Processor
22 Jun 2020
School
Department
Course
Professor
Document Summary
Types of joins: theta join: rows from r and s combine if some condition is true, equi-join: rows from r and s combine if specified attributes match in value, e. g. Theta and equi join repeat tuples with same name, but natural join does not repeat. Indexed nested loop join - use index to avoid searching whole of second table. S tuples for one tuple of r c can be estimated as cost of a single selection on s using the join condition. , rn denote partitions of r tuples: each tuple r r is put in partition ri where i = h(r[joinattrs]), s0, s1. In matching phase, read both relations: br+bs i/os: total cost = 3(br+bs) Also, hash join shown to be highly parallelizable: sort-merge less sensitive to data skew; result is sorted. cost estimate: # of i/os. Issue 1: minimize synchronization => avoid taking latches during execution.
