Using
SAS to Read Disk directories from Windows Systems
It
can be useful on occasion to be able to look in some detail at the files
we accumulate on the hard disks of our PC. For instance:
to identify the largest
files, the ones most likely to be causing disk space problems
to identify true duplicate
files that may be scattered across one or more disk volumes
to identify which
directories occupy the most space
It is not easy to see these
facts using Windows Explorer.
Here are details
of a SAS program that will read directory listings from Windows 95, 98
and NT (I haven't yet seen 2000 or ME) and produce, at present, very
simple reports. Once one has the data, reports can be generated
according to need, anyway. The reports use ODS, but the programs will
work with Version 6 of SAS software if the ODS statements are removed.
The program, as
supplied, requires a directory called 'Diskstats', with subdirectories
'data', 'data sets', 'outputs' and 'programs'. These can all be changed
by modifying the few lines at the start of the code.
The program reads from a text file
generated using the DOS command 'DIR' (from the MS-DOS Prompt[Windows
95]/Command Prompt [Windows NT] in Start Menu>Programs) as follows:
DIR
c:\*.* /s > d:\diskstats\data\sample.txt
The name of the
data file needs to be changed within the code according to the name you
assign it.
The drive letters in the DIR
command will need to be changed according to your own PC setup.
The program has not been designed
as an example of good programming style. Rather, it has been designed as
an exercise for students, to show a number of very useful capabilities
and features of SAS in use.
In reading the
input data, the program first creates a version of the data as a SAS
data set that is very large, and this requires an appropriate amount of
disk capacity - i.e. large, perhaps 30-40 megabytes. This is done on
purpose, to demonstrate how one can easily produce unnecessarily large
data sets. The data set is then deleted in favour of a version created
from it that is, typically, only 10 per cent of the size. This is done
to show how data sets can be made a more reasonable size while
delivering the identical data.
The program also
creates some SAS code and writes it to a file and then %INCLUDEs it.
The final data set contains details of each file and directory for the
disk volume being processed.
