topic Re: Theory of relativity in SAS programming in SAS Programming

Theory of relativity in SAS programming

LeonidBatkhan — Sat, 11 Apr 2020 02:33:16 GMT

"Creativity is intelligence having fun"

~ Albert Einstein

The last time I checked, April Fools' Day has not been cancelled this year. Moreover, we all desperately need it as it is scientifically proven that laughter is an immune booster.

No kidding, laughing is a serious matter. That is why today we are going to get serious, very serious, and tackle the most famous and the least understood scientific theory of all time – the theory of relativity – as it relates to programming in general, and SAS programming in particular.

Fasten your seat belt!

Any computer program can be described in terms of space and time. For example, SAS code originates in our 3-dimensional heads’ space, transforms into a two-dimensional space entity in a Program Editor, and then progresses through compilation to execution in a one-dimensional time line sequence defined by itself and multi-dimensional SAS computational space also known as infrastructure. This introduces a fundamental concept of relativity theory – spacetime as a unified entity of space and time.

Creative use of the spacetime principle opens unlimited possibilities for SAS code developers. For example, you can embed a DATA step within another DATA step or a PROC step and vice versa. You just need to apply the appropriate spacing to ensure your code has proper readability.

The theory of relativity consists of two interrelated parts: special relativity and general relativity.

Special Theory of Relativity

Special relativity applies to all coding methods in the absence of gravity. Consider running the following SAS code:

data AAA;
   set SASHELP.CARS;
   where ORIGIN eq 'Asia';
   data BBB;
      merge AAA SASHELP.CARS;
      by MAKE;
      where TYPE eq 'Truck';
   run;
run;

While it might seem unusual, this is perfectly legitimate code that 1) does not mention gravity and 2) runs without ERRORs. As you can see, the code for data BBB step is totally embedded within the data AAA step, and it even recursively merges its parent data AAA within itself. If you think about it, it’s pure spacetime – a combination of proper code spacing and timing of execution.

Or, you may want to try this code where PROC PRINT is nested within a DATA step and prints its parent data set:

data A;
   x=1;
   y=2;
   proc print data=A;
   run;
run;

You can even further simplify it by removing the data set names:

data;
   x=1;
   y=2;
   proc print;
   run;
run;

It still runs perfectly fine and as expected, even though no data set names are specified.

Here is another beauty where PROC SQL is embedded within a DATA step and recursively reads its own parental data:

data A;
   set SASHELP.CARS;
   where make eq 'Ford';
   proc sql;
      select * from A;
   quit;
run;

The above examples are possible due to coding relativism – the idea that views are relative to differences in perception and consideration. What you see in a SAS program may be different from how the SAS compiler sees and interprets it. Moreover, as you will find below, what you consider as program code could be construed and interpreted as data, and vice versa.

General Theory of Relativity

General relativity applies to objects and methods in the presence of gravity and explains its relation to other objects and methods. It is general relativity that deals with such eminent phenomena as black holes: regions of spacetime exhibiting gravitational acceleration so strong that nothing, I mean, nothing – no objects, particles or even electromagnetic radiation such as light – can escape from it.

The following code example illustrates the concept of a black hole as it relates to SAS programming:

;;;;;;;;;;;;;;;;;;;;;;;;;
; %let time = LINE      ;
; %let s=%length(&TIME) ;
; data B                ;
;    input              ;
;    TYPE = 'Gravity'   ;
;    &time.s&s          ;
;    data A             ;
;       TYPE = 'Light'  ;
;    run                ;
; run                   ;
;;;;;;;;;;;;;;;;;;;;;;;;;

The code is very similar to the ones above it, but as you can see, we did introduce gravity in it. If you dare to run it, you will see that the embedded DATA step (data A) is prominently absent from the SAS LOG, not to mention that it never runs. Typical black hole: we know it exists, but no one can see it in action.

This can only be explained by such a strong gravitational pull of the data set B that nothing can escape from it, precisely the definition of a black hole. Can you provide your explanation?

Conclusion

I hope by now you are convinced that a combination of spacing and timing can produce quite remarkable effects making your SAS code unbelievably elegant and nerdy at the same time. (The degree to which it’s elegant or nerdy is relative to your attitude, perception and consideration.)

I am optimistic that after reading and sifting through these theoretical fundamentals and coding examples you will reach beyond your traditional thinking, discover your fourth dimension and come up with your own code masterpieces that are related to the relativity theory.

Please share your thoughts and ideas below.

Additional resources

April 1, 2019: Dividing by zero with SAS

April 1, 2018: SAS discovers a new planet in the Solar System

April 1, 2017: SAS code to prove Fermat's Last Theorem

Re: Theory of relativity in SAS programming

ChrisNZ — Wed, 01 Apr 2020 09:41:02 GMT

Wow.

My mind is blown.

So much cleverness condensed in such little space!

Almost like a white dwarf.

Thank you!

Re: Theory of relativity in SAS programming

ChrisHemedinger — Wed, 01 Apr 2020 12:04:47 GMT

It's also been established that the faster that you write SAS code -- as your coding activity approaches the speed of light -- the slower that you will age relative to programmers who use other languages.

(Or is it the other way around? I forget things in my old age.)

Re: Theory of relativity in SAS programming

LeonidBatkhan — Wed, 01 Apr 2020 14:14:08 GMT

ChrisHemedinger , the speed of your SAS code writing and your age are irrelevant, as long as you know how to use CALL EXECUTE to write SAS code for you . Or you can even let SAS write batch scripts for you .

Re: Theory of relativity in SAS programming

LeonidBatkhan — Wed, 01 Apr 2020 14:11:10 GMT

Thank you, ChrisNZ !

Here is your white dwarf:

data _;
   if 0;
run;

Re: Theory of relativity in SAS programming

yabwon — Wed, 01 Apr 2020 16:41:46 GMT

Chris,

Wasn't it: "the faster you're coding the slower it's running"? 😉 😉

All the best
Bart

Re: Theory of relativity in SAS programming

KirkLafler — Wed, 01 Apr 2020 18:17:07 GMT

Great post as always, Leonid! Thanks for sharing!

Be safe and stay healthy!

Kirk

Re: Theory of relativity in SAS programming

LeonidBatkhan — Wed, 01 Apr 2020 18:22:35 GMT

Thank you, Kirk. I am glad you like it. Hope you and yours are safe and healthy too!

Re: Theory of relativity in SAS programming

SJSlaughter — Wed, 01 Apr 2020 23:55:18 GMT

Leonid,

Thanks for another creative April 1 post! The laughs are both generally and relatively timely.

Susan

Re: Theory of relativity in SAS programming

LeonidBatkhan — Thu, 02 Apr 2020 02:35:50 GMT

Susan,

Absolutely!

Leonid

Re: Theory of relativity in SAS programming

wrrichar1 — Mon, 06 Apr 2020 18:37:25 GMT

Awesome Leonid! That was funny and enlightening. That's a tough balance to strike.

Re: Theory of relativity in SAS programming

LeonidBatkhan — Mon, 06 Apr 2020 18:47:25 GMT

Thank you, wrrichar1!

Yea, I felt like a tightrope walker balancing over the abyss of reality...

Re: Theory of relativity in SAS programming

MichelleHomes — Fri, 10 Apr 2020 04:46:30 GMT

Great post @LeonidBatkhan!

I smiled at the statement you made "What you see in a SAS program is different from how the SAS compiler sees and interprets it."... when I used to teach SAS programming classes I would emphasise the importance of reading SAS code like the compiler would interpret it. Look for the tokens (or missing tokens) and challenge yourself to see what would happen behind the scenes before you run it. In my book, it distinguishes oneself from "doing" SAS programming to "being" a SAS programmer. Anyone can copy/paste code and try to work it out with debugging/troubleshooting but knowing how the code works and what the code will do before you run it, demonstrates a higher level of understanding and being an efficient SAS programmer.

Re: Theory of relativity in SAS programming

LeonidBatkhan — Fri, 10 Apr 2020 14:34:50 GMT

Thank you, MichelleHomes, I am glad it made you smile. Your point is exactly what I exploited here. 😜

Re: Theory of relativity in SAS programming

sqlGoddess — Mon, 13 Apr 2020 00:17:18 GMT

so beautifully articulated. thank you for sharing these amazing tips on relativity in sas code.

Re: Theory of relativity in SAS programming

LeonidBatkhan — Mon, 13 Apr 2020 14:37:48 GMT

Thank you, charus for your such a nice comment! I am glad you liked it.

Re: Theory of relativity in SAS programming

Kalind_Patel — Wed, 06 May 2020 11:44:38 GMT

Awesome @LeonidBatkhan,
After reading this article, I wish SAS launch Relativity Mode in the future 😀.

Re: Theory of relativity in SAS programming

LeonidBatkhan — Wed, 06 May 2020 14:23:09 GMT

Thank you, Kalind_Patel,

The "relativity mode" has always been in SAS from its inception: just use proper spacing and watch your timing. 😄

Re: Theory of relativity in SAS programming

Johndberglund — Sun, 02 Aug 2020 21:54:21 GMT

Fun post. Blame me not finding it until now on warped spacetime.

After some thought I realized that we can continue this embedding. Here are three steps deep:

data AAA;
set SASHELP.CARS;
where ORIGIN eq 'Asia';
data BBB;
merge AAA SASHELP.CARS;
by MAKE;
where TYPE eq 'Truck';
proc sql;
create table CCC as (select Make, Model, MSRP from BBB);
quit;
run;
run;

I'm reminded of somebody who said

"Every flea has smaller fleas, upon their backs to bite 'em.

And those fleas and smaller fleas, and so ad infinitum."

One further warning - if you use the "Format Code" button on the final SAS code, it ceases working completely. I gather the black hole sucks everything in. (black holes really suck.)

Re: Theory of relativity in SAS programming

LeonidBatkhan — Mon, 03 Aug 2020 15:31:39 GMT

Hi Johndberglund,

You got it!

However, when I applied code formatting (I don't have the Code Format button in my EG so I had to use Edit -> Format Code) to the last code, I got the following:

;;;;;;;;;;;;;;;;;;;;;;;;;
;
%let time = LINE;
;
%let s=%length(&TIME);
;

data B;
	;
	input;
	;
	TYPE = 'Gravity';
	;
	&time.s&s;
	;

data A;
	;
	TYPE = 'Light';
	;
run;
;
run;
;;;;;;;;;;;;;;;;;;;;;;;;;

which worked as fine as before the code formatting. Apparently, it depends on the state you were in when running your code. (I am in Maryland). Or time of day. Or anything else. In any case, it just reaffirms that theory or relativity is closely related to the the Heisenberg's uncertainty principle...