- File Processing vs Database Processing
- and Metadata
File Processing
The concepts involved in File processing are more closely related
to computer programming than database design. The very first efforts in data
storage involved writing to files from a program. If you have every done any
file i/o using a computer program, you have experienced the concept of file
processing.
There are several good links available in todays assignment that should give you
a good understanding of what exactly is involved in File Processing from a data
perspective. Basically we write data to files in a structured way and retrieve
data from the file in the same way. The files are normally sequential (records
must be processed one after the other) so the process can be very time
consumeing on large files.
Several companies still use this form of file processing and its associated
"transaction processing" schemes. First Data, Mutual of Omaha are two that are
pretty prominent.
Basically, in file processing, we create a structure to contain the data, the
structure is designed in such a way that data can be written to memory and read
from memory in a fixed way. These structures are normally referred to as records
when they contain data. Here is a
link to a cobol program that defines a record in the data division. This
program can read data from this structure but no others. As you can see this
program reads data from a StudentFile that has several elements contained in it.
This file is or can be unique to this program. So if another program wants to
write to the file, it may create duplicate data without knowing it. The data is
not always checked for duplicates.
Metadata
As you probably have seen from the links provided, the concept of
metadata is fairly straigt forward. Its simply data about data. In more
simplistic terms, its information that describes the fields of data that we want
to store. For example, if I want to store a name, I need to know how long the
name is or will be and what type of data it is. Most names are collections of
characters, so the metadata for a name field would be something like this:
collection of 25 letters (a string). I might go on to say that the name cannot
be null, and it must contain at least 5 letters. These are all data about data
or information describing the field name. There are several good links available
in todays assignment that should give you a good understanding of what exactly
is involved in metadata
The cobol program we looked at in the last section has a good example. Here is a
link to a cobol program . Notice under the DATA DIVISION the section that
describes the record the program reads. Notice the 'pic' statements after the
fields, these are a form of metadata for the program. They tell the program how
to display the data elements it will read.