WARNING: This page is a sample of news library INTRANET. Some links do not
work and information may not be current.
Tape Loading & Data Analysis
Quite often, databases are acquired using 9-track magnetic tapes. This
method of loading data, which is very common among state agencies, is
regularly used on mainframe computers because large amounts of information
can be stored on tape very cheaply. Depending on the database, a single
acquisition can require anywhere from 1 to 20 tapes.
We can then feed the tape(s) into a 9-track tape machine, and use the
software 9-Track Express to extract the data from the tape and
import it into a FoxPro table, or any other text format. 9-Track can:
- transfer records stored in ASCII (American Standard Code for Information
Interchange) or EBCDIC (Extended Binary Coded Decimal Interchange Code)
on labeled or unlabelled tapes,
- handle unblocked, fixed blocked, variable blocked or variable unblocked
formats,
- translate a variety of number, field types including packed, binary,
COBOL, COMP-3 and zoned decimal formats (However, most formats are provided
in character or number field types).
There are five basic things a reporter should know when acquiring a
database that requires magnetic tapes:
- Record Layout and Code Descriptions (Also called File Description,
File Format, or File Structure)
Without a record layout, it's very difficult to transfer the data into
a readable format. It typically provides the field names, the starting
position of each field, length of each field and field type (i.e. character,
numeric, data, etc.). It should also provide the blocking factor, record
length, and tape density (see below). Using the record layout, we can then
set up a "Profile" or table structure in 9-Track Express that
will eventually extract the data into the desired destination (which is
almost always FoxPro).
A reporter should also ask for any code descriptions. Quite often, many
fields in a database are not self-explanatory, such as name, address, etc.
Tables with codes must be accompanied by some way to determine what these
codes means.
- Blocking Factor
In order to read the data on a tape, the tape must
be moving at a constant speed. But if you want to stop the tape at any
given point, and later resume reading, there needs to be a gap between
records in order for the tape drive to "get up to speed". This
gap is created by a technique called "blocking". A series of
records are grouped together with the gap occurring after every 10 or 20
records. The number of records within each block is called the blocking
factor. Thus, if 10 records with a length of 100 bytes were put in a block,
the blocking factor would be 10 and the size of the block would be 1,000.
This is an example of a fixed block, where a set number of records of equal
length are placed in a block. To learn more about other types of blocking
see David Raynor or Pat Stith.
- Tape Density
Tape Density describes the way in which information is
placed on a computer tape. Basically, the higher the density, the more
data one can put on the tape. A tape density of 6250 BPI (Bits Per Inch)
is typically the density we request. However, if the agency is unable to
provide this type density, our tape machine can read other densities.
- Tape Structure
Ever read a book without a title or table of
contents?
Well, that's what reading a tape without a label is like; it can be done,
but it would be easier if a little more information was provided. A tape
with a IBM Standard Label includes the date the tape was written, the name
of the file the agency has given the tape and a beginning and end tape
mark. Tape marks are simply used to determine where the tape is to begin
and end "reading" or transferring the records. Otherwise, the
computer would continue to read past the file thinking that stray characters
and blank spaces were records. In fact, tape marks are absolutely necessary
if the agency decides to put more than one file on a single tape. The user
needs to know where the first file ends and the second begins.
- Coding
9-Track tapes are based on binary code. While all computers
use a system of "1" and "0", the sequence used to represent
numbers and letters differs. The two most common forms are EBCDIC and ASCII.
Most tapes we acquire are coded in EBCDIC, which is not a problem because
9-Track Express can translate that code into ASCII. So when asking for
a database, a reporter should typically go ahead and request EBCDIC
format.
For more information on how to acquire databases via 9-track
tapes or
more information about the tape loading process, please email David
Raynor or call him at 829-4798.
If you are an N&O reporter or editor, you may send a
request for research to the news
researchers.
If you're not affiliated with The News & Observer
and need research assistance, please contact
Pisys IQ.