topic Re: apply proc freq on a variable in a large dataset in SAS Programming

apply proc freq on a variable in a large dataset

librasonali — Thu, 03 Mar 2022 07:14:19 GMT

Hi,

I see that when I apply proc freq on a variable in a large dataset, the result gets truncated to scientific notation. Eg. for the below on a dataset of 35 Million records

Proc freq data=sales;
Tables year_month ;
Run;

Gives me the frequencies like follows

202108	1.06E+08
202109	1.05E+08
202110	1.02E+08
202111	1.04E+08
202112	1.01E+08
202201	98690488
202202	81642710

I want the numbers to appear completely with a comma separator.

I tried using formats (best12. and comma12.) in the tables statement but to no avail.

Help appreciated! how can i use PROC Template.? not able to figure out syntax

Re: apply proc freq on a variable in a large dataset

Kurt_Bremser — Thu, 03 Mar 2022 07:35:44 GMT

Use the FORMAT option in the TABLES statement:

proc freq data=sales;
tables year_month / format=20.;
run;

Re: apply proc freq on a variable in a large dataset

FreelanceReinh — Thu, 03 Mar 2022 09:35:44 GMT

Hi @librasonali,

I would write the PROC FREQ output to a dataset (OUT= option of the TABLES statement) and then use PROC PRINT or PROC REPORT to display the frequencies (and percentages if needed) in the desired format. Here's an example including the cumulative frequencies and percentages (via the OUTCUM option):

proc freq data=sashelp.heart;
tables smoking_status / out=want outcum missing;
run;

proc print data=want;
format count cum_freq comma12.
       percent cum_pct 6.2;
run;

(The FORMAT= option of the TABLES statement applies to crosstabulations.)

Re: apply proc freq on a variable in a large dataset

ballardw — Thu, 03 Mar 2022 14:44:49 GMT

In addition to @FreelanceReinh for variables that may have lots of values you might consider the NOPRINT option on the Proc Freq statement along with the OUT= to create a data set. No reason to create two output tables if not needed.