Information Technology Reference
In-Depth Information
etc. A plain text file is generally considered to be an unstructured data file, even
if it contains structured data in the earlier sense, because there is nothing in the
file itself that describes how the information is structured. Other types of data
might be log files or mail message stores, and these will usually be text files.
Data collection applications write data either as raw text files or, quite often,
into a relational database.
Unstructured files
Unstructured files are plain text files that are not directly associated with a spe-
cific application. The data in unstructured files may or may not be presented in
a column format. If it is columnar, then there is some convention to define the
columns. It may be that fields (or variables or columns) are separated from
each other by a tab character (called tab separated), or the fields may be sepa-
rated by a comma or other punctuation character (called comma separated), or
the fields may be defined by their fixed width. Fixed-width field files are usu-
ally output by data collection applications, and are somewhat of a left-over
from the old days of punch-card data storage, where data space was very lim-
ited and needed to be used as efficiently as possible.
Often data will be supplied in unstructured text files, but will also have an
associated document, sometimes called a codebook, that describes how the
text is to be interpreted as data. The following example illustrates a typical
codebook. Note, for example, that a “2” in the code column for question #4
means that the participant was married sometime during the months of
April to June. More specifically, the “29” before it in the frequency column
indicates that 29 study participants were married sometime during the
months of April to June. In this way, the codebook gives meaning to the
numbers in a dataset.
Examples: AFDA Study Code book
In Home Questionnaire Code Book II, S.1
Frequency Code
Response
Variable
Type/
Name
Length
15, ask Q.3-5
.
3. Since {MOLE}, did you get married?
If AGE
≥
H2GI3
num 1
12031
0
no [skip to Q.5]
97
1
yes
2610
7
legitimate skip
[
R
<
15 or age not given]
