There are many ways to use /tmp in SAS, indeed almost as many ways as you can use your own subdirectory structure. One example is the use of a large file of raw data to create a SAS data set that will be stored in your own file structure (your own subdirectories). You might ftp the file containing the raw data from another system to /tmp (or to a subdirectory that you create in /tmp), then use the following SAS program to read it:

libname  perm '~/sasdata';
filename  indata  '/tmp/mydir/raw.data';
data perm.project8; infile indata;
input <variable names and specifications>

This would leave the space available within your quota free of the large raw data set while SAS is creating and saving the permanent SAS data set.

One more way of managing limited working space is to ask SAS to clear it out whenever previously-created temporary SAS Data Sets are no longer needed. Some programs create a number of SAS Data Sets, one after the other, leading up to a "final" version that is the one to be used for analyses. All temporary SAS Data Sets, however, are kept in the working space until the end of the SAS session, or until the user asks SAS to delete them, whichever comes first. If you are running out of working space, and you have a number of DATA steps in your program, you can delete any unneeded SAS Data Sets with PROC DATASETS. At the end of the following example, only FINAL remains in the working area. Without the PROC DATASETS step, all four SAS Data Sets would still be saved in the working area.

data one; infile 'raw.data'; input id test1 test2 sex $;
data two; set one; if  sex = 'f';
data three; set two; average = (test1 + test2) / 2;
data final; set three; if  average gt 50;
proc datasets library = work; delete one two three;
run;

The first example shows how to access compressed raw data residing on UNIX disk. The PIPE keyword is used in the FILENAME statement, and the UNIX 'uncompress' command is placed inside the single quotes:

filename test1 pipe 'uncompress -c raw.file.Z';
data cmpress1;
  infile test1;
  input <--variable and format specifications-->

For further information on saving resources while using large amounts of data, see the section above entitled "Using Large Data Sets in the UNIX Environment".

sas -work /tmp1

and SAS will direct all its work files to /tmp1. Note that you must specify alternative workspace when you invoke SAS: workspace cannot be transferred to another place during a SAS session. Note also that if you use a workspace other than the default and your SAS session ends abnormally, it is your responsibility to make sure that any working files left behind are removed to free the space for other users.

If you run out of space during a SAS job and need to estimate whether your job will run if you invoke SAS and specify another directory, first determine how much space is available in the directory (go to the directory and type df . -- that's "df" followed by a space, then a period), then look at the size of the SAS or raw data file you wish to manipulate (this can be done several ways, including the lf -l command). If there is free space equal to twice the size of your file, it should be sufficient, unless you are planning to do a sort, in which case you should have at least three times the size of the file in free space available.