BookmarkSubscribeRSS Feed
Cheesiepoof05
Obsidian | Level 7

I have two queries in series.  In the first query I pull date in, but format it as year quarter ( DATE FORMAT=YYQ. AS Date_Year_Qtr).

The next query I'm trying to sum up the sales and was hoping to group by my new Date_Year_Qtr record.  Apparently since it is just formatted as the year qtr it is still splitting my rows by the actual date itself.  Any way around this?  

 

PROC SQL;
CREATE TABLE WORK.Final_Sum AS
SELECT
Date_Year_Qtr,
SUM(sales) AS sales

FROM WORK.Table_2

GROUP BY Fill_Date_Year_Qtr

;
QUIT;

6 REPLIES 6
PaigeMiller
Diamond | Level 26

To use the formatted YYQ values, you could compute the sums using PROC SUMMARY or PROC MEANS. It won't work in PROC SQL, which will not use the formatted values of the GROUP BY variable to create sums.

 

proc summary data=table_2 nway;
    class fill_date_year_qtr;
    var sales;
    output out=final_sum sum=;
run;

 

In general, I think it is a mistake to resort to PROC SQL for even simple statistics, and learning PROC SUMMARY is very valuable for many reasons. There are occasions when SQL does just as good a job as PROC SUMMARY, but these are few; this is just one example of where PROC SUMMARY works better.

--
Paige Miller
Kurt_Bremser
Super User

You use Date_Year_Qtr in the SELECT, but Fill_Date_Year_Qtr in the GROUP BY. Such will always lead to an automatic remerge, where all observations from the dataset will be kept.

To get a sum for a formatted categorical variable, use PROC MEANS/SUMMARY.

In SQL, you would need to use the PUT function to create a new variable with the format, and use that new variable in the GROUP BY; use no other variable in the SELECT except those that are results of SQL summary functions.

Cheesiepoof05
Obsidian | Level 7

Hi Kurt, that's my fault I was trying to simplify my example and meant to have the select and group by statement the same.  They truly are the same in the real life situation.  Could you give more details around creating a new variable with the put function?

Kurt_Bremser
Super User
proc sql;
create table final_sum as
  select
    put(date_year_qtr,yyq.) as date_year_qtr,
    sum(sales) as sales
  from table_2
  group by calculated date_year_qtr
;
quit;
Tom
Super User Tom
Super User

Convert the values to a single day in that quarter.  For example by using the INTNX() function.

create table final_sum as
select intnx('qtr',Date_Year_Qtr,0) as Date_Year_Qtr format=YYQ6.
     , sum(sales) as sales
from table_2
group by 1
;

Note it would be much simpler to use PROC SUMMARY

proc summary nway data=table_2;
   class date_year_qtr;
   var sales;
   output out=final_sum sum=;
run;

But the actual value of DATE_YEAR_QTR is not guaranteed to be the first day of the quarter.

Cheesiepoof05
Obsidian | Level 7

Thanks all, I'll look around and see if I can find some good tutorials on PROC SUMMARY.

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

Register now!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 6 replies
  • 465 views
  • 0 likes
  • 4 in conversation