Programming with CASL in SAS Viya for Learners – first steps

4 Likes

I often used to joke with younger members of one of the programming teams I worked with that “real programmers don’t use GUIs”. I was only half joking though as while I use GUIs (graphical user interfaces) for programming and I’ve been a big fan of SAS Enterprise Guide and SAS Viya’s visual programming products I still believe that in order to be a really good professional programmer nothing can beat actually learning to hand craft the code yourself rather than relying too much on code generators.

Therefore in this series of articles I shall be demonstrating how to use CASL (Cloud Analytical Services Language) in SAS Viya for Learners using SAS Studio. We will start by seeing how to initialize a CAS session, examine the many sample data sets provided, run a simple statistical process and terminate the session.

What is CASL?

CASL is the scripting language used to execute CAS actions in a CAS server. CAS actions perform tasks and are grouped together in action sets based on common functionality.

Starting a session

Start SAS Viya for Learners and open SAS Studio from the menu.

Create a new SAS program and start a CAS session. This can very simply be done by running the CAS statement and (optionally) naming the session. In this case we will call it CASCONN (if we don’t give it an explicit name the session will, by default, be called CASAUTO).

/* Connect to the cas server and call the connection casconn*/

cas casconn;

Discover which data sets are available

SAS Viya for Learners has a large number of sample data sets preloaded and organized into caslibs (libraries on the CAS server). We can discover what they are by assigning SAS librefs for all the existing caslibs. We can also run Proc Caslib to get details on the data sets held in a specific caslib.

/* Assign SAS librefs for all existing caslibs */

caslib _all_ assign;

/* List the tables in one of the sample caslibs */

proc casutil;
	list tables incaslib="busana4m";
run;

Perform our first analysis

As previously mentioned, CAS actions operate on the data sets in the CAS server and are grouped into action sets. We will use the COUNTRIES data set in the BUSANA4M caslib and run the CORRELATION action in the SIMPLE action set to see if there is any correlation between years of education, UN id and Gross Domestic Product per head. This will produce a matrix showing the Pearson Coefficient for each combination of the variables and is equivalent to running Proc Corr in a non-CAS session. The code is quite simple, although the action set does have many options which you can add to vary the methods used.

/* Perform a simple correlation test */

proc cas;
	simple.correlation / 
		table={name="countries" caslib="busana4m"}
		inputs={"gdp_per_capita" "years_of_education" "un_id"};
run;
quit;

We can see from the results (highlighted in green) that, not surprisingly, there is a strong positive correlation between years of education and GDP per head and a negative correlation between UN ID number and GDP per head.

Terminate the session

Finally, once you have finished your work it is good practice to explicitly terminate your CAS session to free up resources on the server.

/* Terminate the session */

cas casconn terminate;

In future articles we will be taking a deeper look into CAS programming in SAS Viya for Learners but if there are any areas you would like covered or any other questions or comments, please leave a message below.

Quentin · ‎08-15-2022

Sounds like a great series, thanks. Looks like Viya for Learners is only available to people affiliated with a university. Is there a Viya learning environment accessible to independent learners?

ChrisBrooks · ‎08-15-2022

Thanks @Quentin - unfortunately SAS Viya for Learners access is currently limited to those with an accredited university affiliation. Having said that I know SAS is a company which does listen to its customers so if there is enough interest maybe it will be opened up to independent learners in the future.