Prev: Need a shell script to run the SAS Jobs
Next: How to Pump $1,000s in CASH & Checks to your door.
From: JanineJDL on 14 May 2010 03:52
Hi guys.
I’m required to track the outcome of a decision agent on a weekly
basis. Dimensions being customer number, product and decision agent
approved or decline outcome. In order to monitor if decision agent is
stable.
A merge wont work as a customer can move in-between the outcome status
and product offered on a weekly basis. This is a real first one for me
and having a bit a "writers block".
Any help? Suggestions?

JDL
From: Arthur Tabachneck on 15 May 2010 11:41
Janine,

It is difficult to respond without knowing more about what you are
trying to do.

First, example data files would definitely help, preferably in the
form of a data step.
Second, an example data file or output, showing what you want to end
up with would also be essential.

Art
-----------
On May 14, 3:52 am, JanineJDL <janine.dela...(a)standardbank.co.za>
wrote:
> Hi guys.
> I’m required to track the outcome of a decision agent on a weekly
> basis. Dimensions being customer number, product and decision agent
> approved or decline outcome. In order to monitor if decision agent is
> stable.
> A merge wont work as a customer can move in-between the outcome status
> and product offered on a weekly basis. This is a real first one for me
> and having a bit a "writers block".
> Any help? Suggestions?
>
> JDL

From: Arthur Tabachneck on 17 May 2010 08:05
Janine sent me the following off-line:

On May 15, 5:41 pm, Arthur Tabachneck <art...(a)netscape.net> wrote:
> Janine,
>
> It is difficult to respond without knowing more about what you are
> trying to do.
>
> First, example data files would definitely help, preferably in the
> form of a data step.
> Second, an example data file or output, showing what you want to end
> up with would also be essential.
>
> Art
> -----------
> On May 14, 3:52 am, JanineJDL <janine.dela...(a)standardbank.co.za>
> wrote:
>
>
>
> > Hi guys.
> > I’m required to track the outcome of a decision agent on a weekly
> > basis. Dimensions being customer number, product and decision agent
> > approved or decline outcome. In order to monitor if decision agent is
> > stable.
> > A merge wont work as a customer can move in-between the outcome status
> > and product offered on a weekly basis. This is a real first one for me
> > and having a bit a "writers block".
> > Any help? Suggestions?
>
> > JDL- Hide quoted text -
>
> - Show quoted text -

Hi

Have been working on it over the weekend and seem to have come up with
a fairly simple solution.

Data step below:


DATA Prescore.Botswana_vintage_Report (keep= CUS_N RUNDATE
Decision_Cat PRODUCT_TYPE1 LIMIT1 Week1 Week2 Week3 Week4 Week5);
SET Prescore.Botswana_vintage;
format Week1 1.
Week2 1.
Week3 1.
Week4 1.
Week5 1.;
if upcase(compress(RUNDATE))= '2010-02-28'd then Week1 = 1.;
Else Week1 = 0.;
if upcase(compress(RUNDATE))= '2010-03-07'd then Week2 = 1.;
Else Week2 = 0.;
if upcase(compress(RUNDATE))= '2010-03-14'd then Week3 = 1.;
Else Week3 = 0.;
if upcase(compress(RUNDATE))= '2010-03-21'd then Week4 = 1.;
Else Week4 = 0.;
if upcase(compress(RUNDATE))= '2010-03-28'd then Week5 = 1.;
Else Week5 = 0.;
Run;

However SAS log giving me

" ERROR: Invalid date/time/datetime constant
ERROR 77-185: Invalid number conversion"

Outcome Im trying to achieve should be something like this

Customer Product Week1 Week2 Week3 Week4
Janine FTL 1 0 0 1

Simply I reflected in week 1 and week 4 file run.
The second problem I will be having is that I have a few other
dimensions ie product changes, approve or decline

On May 14, 3:52 am, JanineJDL <janine.dela...(a)standardbank.co.za>
wrote:
> Hi guys.
> I’m required to track the outcome of a decision agent on a weekly
> basis. Dimensions being customer number, product and decision agent
> approved or decline outcome. In order to monitor if decision agent is
> stable.
> A merge wont work as a customer can move in-between the outcome status
> and product offered on a weekly basis. This is a real first one for me
> and having a bit a "writers block".
> Any help? Suggestions?
>
> JDL

Janine,

I've posted your response to SAS-L as sending questions to individual
SAS-Lers defeats two of the major benefits of SAS-L: collaboration and
historical retrieval.

I don't have time to look at your code this morning, but did notice a
couple of things. First, in using a date constant, you have to use a
valid one. Try "28FEB2010"d instead of "2010-02-28"d, 7MAR2010"d
instead of "2010-03-07"d, etc. However, I saw a related inconsistency
in your code, namely that you were using the compress function on
rundate which you were then comparing with the date constant. Is
rundate a character or date variable?

Art
 | 
Pages: 1
Prev: Need a shell script to run the SAS Jobs
Next: How to Pump $1,000s in CASH & Checks to your door.