Data Management in SAS/FSP

By Ying-Hua Liu

The SAS statistical program offers many supplemental tools, such as SAS/AF and SAS/SQL, that are very useful to intermediate and advanced programmers. Another such product is SAS/FSP (Full Screen Products), an advanced SAS tool used to interactively update and maintain databases. Procedures in SAS/FSP are developed by the Display Products Division. The product allows researchers to create "views in SAS that are operating system-independent. Product development features include design, programming, debugging, support, and preliminary documentation.

Using SAS/FSP software you can:

• browse and edit the contents of SAS data sets;
• enter data into existing SAS data sets;
• create new SAS data sets;
• browse and edit SAS data views created with SAS/ACCESS;
• browse SAS data views created with the SQL procedure in Base SAS software;
• create, edit, and print form letters and reports;
• build and customize end-user applications.[1]

Figure 1. A sample data set viewed in SAS/FSP's full screen mode.

SAS/FSP'S Main Procedures

The four main procedures in SAS/ FSP are: fsbrowse, fsedit, fsview, and fsletter.

First, in order to view a full screen, you must create a new folder in the C:\ drive of your computer and give it a name (e.g., "London). If you forget to create this new location and then refer to it in your code (as described below), all of your work will be put into the SAS/FSP temporary work library, which will be deleted when you close your session.

In SAS/FSP, enter the following statement:[2]

libname librif "C:\London;
run;
proc append base=librif.company
data=sashelp.company;
run;
(This appends a copy of the file company within sashelp to your library librif.)

proc contents data =librif.
company;
run ;

(This displays how many variables there are in company, their types, and any observations in the data.)

Figure 2. Cases can be browsed in the View menu.

Fsbrowse

The fsbrowse procedure provides a custom display in which you can specify how your data are presented and add descriptive text, if you wish. Fsbrowse allows you to browse the contents of a SAS data set one observation at a time.

proc fsbrowse data=librif.
company;
run;

You will now see a full screen, as shown in figure 1. In full screen mode, you can browse each case from the View pull-down menu (see figure 2). More important, you can also find a case using the Search menu. To do so, type the variable, then enter the value or character in which you are interested. For example, if you want to find level5=Steffen Graff, enter level5 as the variable name, then enter "Steffen Graff as the value; since level5 is a string, you should include quotation marks around "Steffen Graff."

Fsedit

The fsedit procedure lets you update the database by editing, deleting, or adding a new observation. To use fsedit, go to the Edit menu and select Update, or write the following fsedit statement. (Be sure to close the fsbrowse window first.)

proc fsedit data=librif.
company;
run;

To edit and delete an observation, first find the observation using the Search menu, then go to the observation. From there, you can change the value for each observation or add and delete values or cases by going to the Edit menu and selecting "Add Record" or "Delete Record."

Figure 3. A sample data set as it appears in fsview (fictitious data is used here).

Fsview

Fsview enables you to browse or edit a SAS data set, displaying it as a table of rows and columns. You can use this procedure to create a new SAS data set. Fsview also provides tools for customizing an fsview application. For example, you can redesign the display by changing the size, position, and colors of the fsview window. In addition, you can add computed variables, which display values that are calculated from other variables in the data set.

For fsview, enter:

proc fsview data=librif.
company;
run;

The output will be similar to the sample shown in figure 3.

Fsletter

The fsletter procedure enables you to create, edit, and print letters and other documents. When creating and editing documents in the fsletter window, you can use all the features of the SAS text editor, including the spell checker.

Keep in mind that your fsletter documents can include named fields. When the document is printed, you can fill in these fields manually, or the procedure can fill in the fields automatically using values from a SAS data set. This is convenient for creating and maintaining form letters, for example. You can print individual copies of the document, or you can automatically print a copy for every observation in a data set. For more information about fsletter, visit http://support.sas.com/91doc/docMainpage.jsp.

Data Selection and Data Merge

Say that you have finished your data entry, and now you want to separate your data into subsets and then examine a single subset. For instance, you might only want information about New York, Tokyo, or London. You can get each data subset using the "where statement:

Data librif.Company;
Set librif.company;
Where level2=Company;
run;

Data librif.Newyork;
Set librif.company;
Where level2=New York;
run;

Data librif.London;
Set librif.company;
Where level2=London;
run;

data librif.Tokyo;
Set librif.company;
Where level2=Tokyo;
run;

Using the "Explorer window, you will be able to see the four data sets you have created. You may now work on all four data sets at once (see figure 4).

Figure 4. Our sample data sets.

Overview

SAS/FSP is a powerful tool for managing databases. Clinical and business research projects with anticipated long periods of data input and management could benefit from this tool. All computers in the ITS computer labs (Macintosh and Windows) offer SPSS software, and all Windows computers in all of the labs offer SAS. If you have any comments and questions about this software, attend one of ITS' free workshops (http://www.nyu.edu/its/classes/) or contact Frank LoPresti at frank.lopresti@nyu.edu.

To receive email notices about statistical package training and clinics, join the "Statistics Forum found in the Forums channel in the Home tab of NYUHome (http://home.nyu.edu).

Footnotes

Author Biography

Ying-Hua Liu is a graduate student at NYU's Robert F. Wagner Graduate School of Public Service and works as a statistician assistant for ITS Faculty Technology Services.