Date: Mon, 18 Jan 2010 09:46:17 -0500
Reply-To: saslearn chicago <sasswamy@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: saslearn chicago <sasswamy@GMAIL.COM>
Subject: Fwd: PREPORT - Date / Week
In-Reply-To: <201001152022.o0FHgMWg001175@malibu.cc.uga.edu>
Content-Type: text/plain; charset=windows-1252
PLEASE Disacard this email , I have found a way for this Report,
thx and sorry for the confusion,
- swamy
---------- Forwarded message ----------
From: SAS Swamy <sasswamy@gmail.com>
Date: Fri, Jan 15, 2010 at 3:22 PM
Subject: PREPORT - Date / Week
To: SAS-L@listserv.uga.edu
Cc: SAS Swamy <sasswamy@gmail.com>
Hello everyone,
I need suggestion in the following report :
Week Rec Cum rec Tot Rec Cum Percent
0 3 3 3 100%
0 3 3 7 43%
1 4 7 7 100%
0 3 3 12 30%
1 4 7 12 60%
2 5 12 12 100%
0 3 3 18 10%
1 4 7 18 20%
2 5 12 18 60%
3 6 18 18 100%
I will be calculating the Week (2nd column), based on 2 date fields for
each rows from the Input file , I would apply rule date2 – date1 / 7 to
arrive the week nbr, later group by week , So I will be getting one row
for one Week number,
Let say , we are working on a date 01/01/2005 ,
week 0 will be 01/01/2005 – 01/07/2005
week 1 will be 01/01/2005 – 01/08/2005
Based on the above report , you eventually will be having only 4 rows
after grouping , Week 0 , 1 , 2 ,3
The challenge is to stack with different break up’s 0 , 0 & 1 , 0 & 1 & 2
…
etc
What would be the best way for this report , I have not used PROC report
yet in SAS ?
here is the steps I am working on , ( may be correct , but longer )
I will group the Week 0 to a data set week_0 , then group the week 1 to a
data set week_1 , this way I would have week_0 ….week_12 ( or whatever the
max week calculated by above formula )
Then combined based on a loop variable for the counter AND data sets
name ,
for the first break up will set week_0 , for the 2nd break up I would set
week_0 , week_1
In the 2nd break up , since the week_1 will have the different cumulative
record count as of Week_1 , I need to replace with the cumulative record
count of week_0 to match both.
same way 3rd break up I would take the cumulative count from week_2 and
replace for week_0 and week_1 …
Any suggestion , Idea for better logic ,Please Advice
- Swamy
| 
