Date:         Thu, 21 Sep 2006 13:23:51 -0400
Reply-To:     Scott Czepiel <sczepiel@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Scott Czepiel <sczepiel@GMAIL.COM>
Subject:      Re: problem with eliminating duplicate observations
In-Reply-To:  <1739703E99A04F4E9354B69912A9D5D70E0EAF51@EXCH2K3.india.birlasoft.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

On 9/21/06, Bhavani Shankar <bhavani.shankar@india.birlasoft.com> wrote:
>
> 2. Now I want month wise  Unique count of card account number ( Proc
> sort with Nodupkey will not solve my problem)


 This part by itself can be easily solved using proc sql as you mentioned in
step 1.

select distinct month, cardnum, count(*) as count from dataset group by 1,
2;

It's the "group by 1, 2" clause that ensures you will get unique card
numbers within each value of month.

I want these result of 1 and 2  in a single data set
>

I do not think you can include aggregate totals (from step 1) in the same
dataset as that containing unique respondents per month (as in the example
above).