relationaldb by fP7kk9YK


									Relational Databases

      Alex Wood
        Why Relational Databases?

   Originally, Databases Were Stored as “Flat Files”
   Flat Files Are Very Poor at Handling Queries
   Breaking Data into Relational Models:
       Increases Flexibility
       Improves Query Performance
       Uses Less Space (Avoiding Excessive “NULL” Fields)‫‏‬
                Flat Files Are Evil

Number, Name, Position, Age,           Flat File DBs are OK
  College|27, Jordan Babineaux,
  SS, 27, Southern Arkansas|83,         for Storage of Data
  Deion Branch, WR, 30,                 Unlikely to Be
  Louisville|79, Red Bryant, DT,
  25, Texas A&M|81, Nate                Significantly Modified
  Burleson, WR, 28, Nevada|11,         Adding New Columns
  Deon Butler, WR, 23, Penn
  State|89, John Carlson, TE, 25,       is Difficult
  Notre Dame|59, Aaron Curry, LB,      How Do You Model a
  23, Wake Forest
                                        List of an Author's
                                        Books in a Flat DB?
A Few Tables
     A Few Tables

How do we model relationships?
Relation Tables

           Using only shared
            keys, facilitates easy
            linking of tables in
           Allows for modular
            data storage.
           Can add attributes to
            further describe
Deriving Interesting Information

                    How many students
                     from a given school
                     does a company hire?
                    What percent of
                     employees come from
                     that school?
                    What is the average
                     commute distance of a
                     company's employees?
                  Our Project

   The CTPP Project involves storing game state and
    game results for an educational video game.
   We are able to leverage Relational Databases to
    efficiently store game data, and build queries to
    obtain interesting information for post-game
   Dealing with changing requirements for stored
    data is a manageable obstacle with this

To top