Introduction to Data Management

An Overview of Data Management
15873 Tandem Computers Incorporated 1-5
Figure 1-3. Tabular Structure in a Relational Database
S8020-003
311
310
255
110
233
340
105
260
380
123
250
360
10
10
9
5
9
10
5
7
10
6
9
10
Roberts
Roberts
Aaron
Crandall
Daen
Roberts
Crandall
Mandell
Roberts
Crandall
Aaron
Roberts
Employee Extension Office No. Dept. No.
Manager
Table
Column
Row
Adams, Joan
Anderson, Samuel
Archer, Edward
Baker, Susan
Belding, Frances
Bellamy, Roger
Blanchard, Ken
Brent, Moris
Cash, Roselle
Catlett, Mary
Chatterly, Irma
Clancey, Donald
6277
7728
9962
6301
7544
6276
5760
7872
2798
9874
2267
9230
Because files in a relational database are written in a uniform format, they can be
manipulated by powerful aggregate operators that let users operate on all rows in an
entire table. Users can sort rows by selected fields, join two or more tables to produce
another table, group rows by a particular criterion, and display items in the database
in many other ways. Users can select particular rows through a series of
specifications. They can choose columns simply by assigning names to them.
Because files in a relational database are physically independent (although logically
interdependent), they can be conveniently expanded. New files are easy to add, and
existing files can be redistributed over a network without structural changes to other
files in the database. This can all be done without the costly program redesign,
recoding, or recompilation that would otherwise be necessary to compensate for
expanded or redistributed databases.
Furthermore, relational systems eliminate the need for a geographically distributed
application to work with several independent databases that each reside at different
locations. Instead, these applications can work with a single database that is itself
distributed over the network.