THE DATABASE APPROACH - Database Design and Development: An Essential Guide for IT Professionals - page 14

Such variations are likely causes for data integrity problems. Variations in names
and addresses may cause erroneous printing of names and addresses on documents
mailed to customers.
Inflexibility When an application cannot adapt to changing requirements, we say
that the application is inflexible. By the very nature of sequential files, a file-
oriented system can process transactions only in batch mode. That is, transactions
are accumulated and processed as batches. You cannot print an invoice for a single
customer for a single sale. Sequential files allow retrieval of data records, one at a
time, in sequence. Such files do not possess the flexibility to meet new and chang-
ing information requirements.
Suppose you are interested in finding and listing all the purchases made by cus-
tomers in Japan for the past three months. Or you want a list of all customers in a
certain zip code range. It is very difficult to satisfy such ad hoc queries in a file-
oriented system without reprogramming.
Limited Data Sharing Consider two typical applications, namely, order process-
ing and inventory control. Each of these applications needs data on products. But
data on products are repeated in each of these two applications when the business
depends on file-oriented data systems. There is no sharing of data on products
between the two applications. If product descriptions of certain products are
changed, these changes must be made in both applications.
Difficult Data Integration Let us get back to the bank example with separate cus-
tomer files in the checking, savings, and loan applications. If you wanted to combine
data from these applications and send consolidated statements to customers
showing the transactions in all three accounts, it would be nearly impossible with
file-oriented data systems. You would have to run special programs to extract
banking transactions from each application. Then you would have to come up with
methods for matching customer accounts from each application and consolidate the
transactions in a single statement, which is not an easy task. The proliferation of
files and duplication of data continue as each new application is implemented with
its own set of files.
Poor Enforcement of Standards and Controls Standards relate to data names,
formats, value restrictions, and access controls. When duplicated data are spread
across many applications, it is extremely difficult to enforce standards. For example,
if the standard in your company is for the customer name field to be 35 bytes,
then you will have to impose this standard, not in one place, but in many applica-
tions that store customer names. Suppose you have to include a business rule that
the employee daily wage rate must be between 0 and 100; you may have to stipu-
late this rule in at least two different applications, namely, payroll and human
resources.
Problems with the resolution of homonyms and synonyms deserve special atten-
tion. File-oriented data systems are likely to have problems with these.
Homonyms. If a single field name represents two or more objects in different appli-
cations, it is called a homonym. For example, the field name balance may represent