topic How to take highest values from different variables and calculate mean in SAS Programming

How to take highest values from different variables and calculate mean

Skillside — Sat, 14 Mar 2020 21:38:43 GMT

Hi,

this time i would like to calculate mean value from different variables (their highest value). Please, suggest a solution.

data have;
 input ID age1 age2 age3 age4 age5 ;
 cards;1 1 3 6 9 2
2 6 2 3 4 4
3 7 2 9 0 0;
run;

the mean should be calculated from 9, 6 and 9

Thank You

Re: How to take highest values from different variables and calculate mean

Kurt_Bremser — Sat, 14 Mar 2020 21:51:06 GMT

Use

max(of age:)

in a data step to create the max for each observation, then calculate the mean as usual with sql or proc means.

Re: How to take highest values from different variables and calculate mean

Skillside — Sat, 14 Mar 2020 22:09:48 GMT

that was too easy 😉

thank you

Re: How to take highest values from different variables and calculate mean

novinosrin — Sat, 14 Mar 2020 22:28:40 GMT

data have;
 input ID age1 age2 age3 age4 age5 ;
 cards;
1 1 3 6 9 2
2 6 2 3 4 4
3 7 2 9 0 0
;
run;

data mean_of_max_age;
 _n_= 0;
 do until(z);
  set have end=z;
  max_age= max(of age1-age5);
  _n_=sum(_n_,n(max_age));
  _iorc_=sum( _iorc_ ,max_age);
 end;
 Mean= _iorc_/	_n_;
 drop age1-age5 id;
run;

Re: How to take highest values from different variables and calculate mean

ballardw — Mon, 16 Mar 2020 19:53:03 GMT

@Skillside wrote:

Hi,

this time i would like to calculate mean value from different variables (their highest value). Please, suggest a solution.
data have;
 input ID age1 age2 age3 age4 age5 ;
 cards;1 1 3 6 9 2
2 6 2 3 4 4
3 7 2 9 0 0;
run;
the mean should be calculated from 9, 6 and 9

Thank You

A hint for future example data sets: If you are going to reference the required values it would be a better idea to have those values unique or explicitly state which value from which record. Max is relatively simple but a more complicated question with both 6 and 9 repeated in different records might be another story.