Creating a formula column based on available information

[ilikesas]

Hi,

suppose I have the following equation:  R = A*B, and B = C/D

and suppose I have the following table:

A	B	C	D
a1	b1
a2		c2	d2

In the first line I can directly calculate R1 = a1*b1, but in the second line I need to  R2 = a2*c2/d2,

so how is it possible to write a code that will choose the right formula depending on the available information?

Thank you

ps: in case all variables are available, choose the "straightest" method, i.e  R=A*B

[stat_sas · Accepted solution]

Hi,

Try this.

data have;

input A B C D;

datalines;

2 5 . .

1 . 10 5

;

data want;

set have;

R=a*coalesce(b,c/d);

run;

[ilikesas]

Hi stat@sas, ran your code and it was good, thank you!

Could you please also show me how to achieve the same result using the IF THEN ELSE procedures because if my actual case will have more variables and more complex equations I guess it will be too messy to put everything into a single COALESCE?

Thank you

[stat_sas]

Hi,

Below is an example using IF THEN ELSE statement.

data want;

set have;

if b ne . then R=a*b;

else R=a*c/d;

run;

[Reeza]

Since you'll end up hard coding things not necessarily. If you have multiple formulas/conditions I'd probably calculate them all, but then use a coalesce to determine which is the final one to keep. The issue then becomes how to keep your log clean, if that's a requirement.