Data Conversion: Doing it Right the First Time - High-Performance Web Databases: Design, Development, and Deployment - page 585

Chapter 41

Data Conversion:
Doing it Right the
First Time

Michael Zimmer

W ,

they usually
do not start with a clean slate. Often, they are replacing an existing appli-
cation. They must always determine if the existing information should be
preserved. Usually the older information is transferred to the new system
— a process known as data conversion.

HEN SYSTEMS DEVELOPERS BUILD INFORMATION SYSTEMS

Data conversion can involve moving data from flat file systems to rela-
tional database management systems (RDBMS). It also can involve chang-
ing from systems with loose constraints to new systems with tight
constraints.

This chapter focuses on laying the groundwork for successfully execut-
ing a data conversion effort the first time around. It is assumed in this chap-
ter that data modeling is being done and that relational database
technology is employed. At the logical level, the terms

entity set, entity, and

attribute are used in place of the terms file, record, and field . At the physical

level, the terms table, row, and column are used instead of file, record, and

field . The members of IS engaged in the data conversion effort are referred

to as the data conversion team (DCT).

COMMON PROBLEMS WITH DATA

The difficulties of a data conversion effort almost always are underesti-
mated. Usually the conversion costs many times more than originally an-
ticipated. This is invariably the result of an inadequate understanding of
the cost and effort required to correct errors in the data. Usually the qual-
ity of the existing data is much worse than the users and development team
anticipate.