outputing more than one variable in proc freq

[Banke]

Hello,

i was asked to print obervations with 5 highest frequencies from a big dataset. i did this to order the frequency in descending order.

PROC FREQ DATA = WEEK6.THCICSAMP ORDER = FREQ;
TABLES ADMITTING_DIAG PRINC_DIAG_CODE;
RUN;

but i want to print , so i tried to create a new data set with only my frequencies for the 2 variables, sort by decreasing frequencies and use proc print :

PROC FREQ DATA = WEEK6.THCICSAMP STACKODSOUTPUT NOPRINT;
TABLES ADMITTING_DIAG PRINC_DIAG_CODE /OUT = WEEK6.FREQDIAG;
RUN;

i noticed that the new data set week6.freqdiaq i created has only one of the variable, PRINC_DIAG_CODE. how can i get all my variables out into the new dataset please?

[Ksharp · Accepted solution]

You need two TABLES statements ,and combine them later by data step.

PROC FREQ DATA = WEEK6.THCICSAMP NOPRINT;
TABLES ADMITTING_DIAG /OUT = WEEK6.ADMITTING_DIAG ;
TABLES PRINC_DIAG_CODE /OUT = WEEK6.PRINC_DIAG_CODE ;
RUN;

[Banke]

Thank you so much, I wanted to do it all in one step.

[Watts]

You could use ODS.  

PROC FREQ DATA = WEEK6.THCICSAMP ORDER = FREQ;
TABLES ADMITTING_DIAG PRINC_DIAG_CODE;
ods output onewayfreqs=out;
RUN;

[Banke]

When  I wanted to print out the observations though, I noticed the PRINC_DIAG_CODE variable is blank

[Banke]

the variables in the two dataset i have created have similar names, so if i use merge command in my datastep, one will overwrite the other, do i have to rename the observations? they both have count and percentage of frequency count in common as variable names. do i also have to create a common variable like ID number to merge them ?

thanks

[Ksharp]

yes. you need rename it before set them all.

proc freq data=sashelp.class noprint ;
table weight/out=weight(rename=(weight=vname));
table age/out=age(rename=(age=vname));
run;

data want;
 set weight age indsname=dsn;
 dsname=dsn;
run;