CMPS 340 Fall 2008
Test #1 Study Guide

• Disks and Data Organization (Chapter 13, Sections 1-7, 10)
  • Double Buffering: be able to explain what benefits it offers over single buffering. Also, given the ratio between the speed at which a program "consumes" the data in an (input) buffer and the speed at which a buffer is filled (by the I/O system), estimate what percentage of the time the program is waiting for an input buffer to fill up. (See Figure 13.4 in Elmasri & Navathe.)
  • Given a file's size (in terms of number of buckets it occupies) and disk drive characteristics (average seek time, etc.), calculate the time needed to perform a binary search in the file (which is just the expected number of bucket fetches multiplied by the time to fetch a bucket).

  • Given a description of a file (number of records, length of records), a distribution of file operations (e.g., 99% single-record fetches, 1% exhaustive sequential reads), and disk drive characteristics (e.g., sector size, average seek time, sector transfer time, etc.), calculate the optimal size for buckets.

  • Given a description of how records and metadata in a bucket are laid out, as well as a description of the particular records in a bucket (e.g., Name: John Gorn, Birthdate: 5/12/1987, etc.), show the contents of the bucket. Also supply code for calculating the location, within a bucket, of the byte at which a specified record begins.

• (Flat) Indexes (Chapter 14, section 1 and sub-section 5.2)
  • Know the differences between primary, clustering, and secondary indexes, and be able to show an index corresponding to given data in blocks/buckets.
  • Explain the tradeoffs between using logical pointers as the "reference" values in an index versus using physical pointers.
• Sequential File Update
  • Given a master file and a transaction file, be able to "manually execute" the Balanced Line Algorithm in order to generate the updated master file.
  • How would you implement the unGet() operation that's used in the Balanced Line Algorithm.
  • cosequential processing: be able to adapt the classic algorithm for merging ordered sequences to solve a similar problem.
• Relational Database (Chapter 3, sections 1-7, Section 7.1, Sections 5.1-5.2, Sections 6.1-6.5)
  • Given an ER diagram, indicate what relations/tables should be in the database, and what their attributes/columns should be.
  • Given tables in a database (with tuples shown) and a relational algebra query, indicate the resulting table.
  • Given schemas of tables in a database (a description of which attributes/columns they have) and a query given in English, give an equivalent query in relational algebra.