Commodore 1541-II - Chapter 6: Relative Data Files

CHAPTER 6
RELATIVE DATA FILES
THE VALUE OF RELATIVE ACCESS
Sequential files are very useful when you're just working with a continuous stream of
data— i.e., information that can be read or written all at once. However, sequential files
are not useful or desirable in some situations. For example, after writing a large list of
mail labels, you wouldn't want to have to re-read the entire list each time you need a
person's record. Instead, you need some kind of random access, a way to get to a
particular label in your file without having to read through all those preceding it first.
As an example, compare a record turntable with a cassette recorder. You have to
listen to a cassette from beginning to end, but a turntable needle can be picked up at any
time, and instantly moved to any spot on the record. Your disk drive works like a turntable
in that respect. In this chapter we will learn about a type of file that reflects this flexibility.
Actually, two different types of random access files may be used on Commodore
disk drives: relative files and random files. Relative files are much more convenient for
most data handling operations, but true random access file commands are also available to
advanced users, and will be discussed in the next chapter.
FILES, RECORDS, AND FIELDS
When learning about sequential files, we did not worry about the organization of data
within a file, so long as the variables used to write the file matched up properly with those
which read it back into the computer. But in order for relative access to work, we need a
more structured and predictable environment for our data.
The structure we will use is similar to that used in the traditional filing cabinet. In a
traditional office, all customer records might be kept in a single file cabinet. Within this
file, each customer has a personal record in a file folder with their name on it, that
contains everything the office knows about that person. Likewise, within each file folder,
there may be many small slips of paper, each containing one bit of information about that
customer, such as a home phone number, or the date of the most recent purchase.
In a computerized office, the file cabinet is gone, but the concept of a file containing
all the information about a group or topic remains. The file folders are gone too, but the
notion of subdividing the file into individual records remains. The slips of paper within
the personal records are gone too, replaced by subdivisions within the records, called
fields. Each field is large enough to hold one piece of information about one record in the
file. Thus, within each file there are many records, and within each record there are
typically many fields.
A relative file takes care of organizing the records for you, numbering them from 1 to
whatever, by ones, but the fields are up to you to organize. Each record will be of the
same size, but the 1541 won't insist that they all be divided the same way. On the other
hand, they normally will all be subdivided the same way, and if it can be known in
advance exactly where each field starts within each record, there are even fast ways to
access a desired field within a record without reading through the other fields. As all of
this implies, access speed is a primary reason for putting information into a relative disk
file. Some well-written relative file programs are able to find and read the record of one
'55