Receiving this NOTE: Invalid argument(s) to the exponential operator "**" at line x column xx.

[sah_biostat]

Hello Team,

I am deriving one variable with exponential operator and receiving this note.How can we rectify this ?

data lb;
set _lb;
GFR_CKEPI = 101 * min((LBSTRESN/kappa),1)**alpha * max((LBSTRESN/kappa),1)** -1*(3.209) * 0.88**AgeAtLB * SEX_factor2 * RACEO; 
run; 

[Quentin · Accepted solution]

Do you maybe have 0**0 or an overflow problem? They would both cause that note.

1    data _null_ ;
2      x=0**0 ;
NOTE: Invalid argument(s) to the exponential operator "**".
3    run ;


4
5    data _null_ ;
6      x=2**10000 ;
NOTE: Invalid argument(s) to the exponential operator "**".
7    run ;

[sah_biostat]

I guess this is due to big constant number.Can we rectify this?

[Quentin]

There is a limit on the largest integer, that varies by OS:

13   data _null_ ;
14     x=constant('big') ;
15     put x= ;
16   run ;

x=1.797693E308

If your number is bigger that, I think you're out of luck.  See e.g.:

https://documentation.sas.com/doc/en/lrcon/9.4/p0ji1unv6thm0dn1gp4t01a1u0g6.htm

https://communities.sas.com/t5/SAS-Programming/quot-Largest-Integer-Represented-Exactly-quot-NOT-Exa...

[PaigeMiller]

---

@sah_biostat wrote:

I am deriving one variable with exponential operator and receiving this note.How can we rectify this ?

data lb;
set _lb;
GFR_CKEPI = 101 * min((LBSTRESN/kappa),1)**alpha * max((LBSTRESN/kappa),1)** -1*(3.209) * 0.88**AgeAtLB * SEX_factor2 * RACEO; 
run; 

---

Please, from now on when you have errors in the log, show us the ENTIRE log for this step (DATA step or PROC).

[ballardw]

What do your values of ALPHA look like?

Did you generate this equation with a SAS model procedure? If so, maybe this could be done as scoring data using the information from the model.

[s_lassen]

You most probably have an overflow error. The actual final value may not overflow, though.

One possibility is to calculate the logarithm of what you want first:

log_GFR_CKEPI = log(101) +log(min((LBSTRESN/kappa),1))*alpha -log(max((LBSTRESN/kappa),1))+log(3.209) +log(0.88)*AgeAtLB +log(SEX_factor2)+ log(RACEO);
GFR_CKEPI=exp(log_GFR_CKEPI);

 That way, you can do the calculation, even if there is an overflow along the way - you will then only get an error message from the EXP function if the final value overflows.