This has to be a really simple question but I have been looking through examples and cannot find a good one. Here is my input:
DATA LISTINP;
infile "~/Pgm1ExtInp.txt";
INPUT ID HEIGHT WEIGHT GENDER $ AGE;
RUN;
I want to sum HEIGHT and WEIGHT as columns in the resulting table so the result would look something like:
GENDER HEIGHT WEIGHT
SUM SUM
Male 210 650
Female 190 600
I am trying to use the PROC TABULATE but it is not working. There seem to be at least a hundred examples of how I could break up HEIGHT by GENDER across HEIGHT values and the same for WEIGHT but none where I can just get those values as columns. If this were a SQL statement it would be:
SELECT GENDER, SUM(HEIGHT), SUM(WEIGHT)
FROM LISTINP
GROUP BY GENDER
ORDER BY GENDER;
I have to think the answer looks something like:
PROC TABULATE DATA=LISTINP;
TITLE 'Summary of Height and Weight';
class GENDER;
var HEIGHT, WEIGHT;
table GENDER , HEIGHT * WEIGHT;
But that doesn't work. All help appreciated. Thanks.
RUN;
PROC TABULATE DATA=LISTINP;
TITLE 'Summary of Height and Weight';
class GENDER;
var HEIGHT WEIGHT;
table GENDER , (HEIGHT WEIGHT)*sum;
run;
Or: height*sum weight*sum; the statistic isn't formally needed as defaults to sum for VAR but I prefer to explicitly state the statistic so I can debug easier.
Your request with height*weight is going to attempt to nest (cross) height and weight.
Also, remove the comma from the VAR statement.
PROC TABULATE DATA=LISTINP;
TITLE 'Summary of Height and Weight';
class GENDER;
var HEIGHT WEIGHT;
table GENDER , (HEIGHT WEIGHT)*sum;
run;
Or: height*sum weight*sum; the statistic isn't formally needed as defaults to sum for VAR but I prefer to explicitly state the statistic so I can debug easier.
Your request with height*weight is going to attempt to nest (cross) height and weight.
Also, remove the comma from the VAR statement.
Thank You! That does it. You will probably be seeing a number of simple questions from me disguised as discussions soon. Thanks again.
If they are questions please post them as questions, Even simple ones. Discussions can't get marked as answered. If a question has been adequately answered then the rest of us, unless curious, know that additional comment is not required and we can spend time on the items that need help.
It's finally time to hack! Remember to visit the SAS Hacker's Hub regularly for news and updates.
Learn the difference between classical and Bayesian statistical approaches and see a few PROC examples to perform Bayesian analysis in this video.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.