Date:         Thu, 4 Mar 2004 11:15:31 -0500
Reply-To:     diskin.dennis@KENDLE.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Dennis Diskin <diskin.dennis@KENDLE.COM>
Subject:      Re: Question about "row-level" processing for a dataset
Comments: To: Sober Hi <soberhi@YAHOO.COM>
Content-Type: text/plain; charset="us-ascii"

Sober,

Here's one way:

data prn_exp;
merge tuition sch_ship(in=_ss);
by person;
/* in case there was no SS for this person */
if not _ss then total = 0;

/* amount from scholarsip this year */
pay_ss = min(cost , total);
cost=cost-pay_ss;
total = total-pay_ss;
run;




HTH,
Dennis Diskin




Sober Hi <soberhi@YAHOO.COM>
Sent by: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
03/04/2004 10:43 AM
Please respond to Sober Hi


        To:     SAS-L@LISTSERV.UGA.EDU
        cc:
        Subject:        Question about "row-level" processing for a dataset


I am fairly new to SAS so be gently please. I am using SAS 6.12 .

I am trying to figure out how to handle "row-level" processing within
a SAS dataset. I think in Oracle/SQL Server, the idea is called a
cursor. Perhaps the approach in SAS would be similar.

Here's my scenario. I have the following two datasets (we have them
already from elsewhere with the following data).

Dataset TUITION:
PERSON  YEAR    COST
DAVE    2004    25
DAVE    2005    30
DAVE    2006    35
DAVE    2007    35
DAVE    2008    40
JOHN    2004    10
JOHN    2005    20
JOHN    2006    25
KIM     2004    15
KIM     2005    10
KIM     2006    0
KIM     2007    25
KIM     2008    50



Dataset SCH_SHIP:
PERSON  TOTAL
DAVE    65
JOHN    75
KIM     80


The relationship between the two datasets:
Dave has received $65 in scholarship money (can only be used to reduce
his tuition expenses). This is found in the SCH_SHIP dataset. His
future tuition expenses are in the TUITION dataset.

So Dave's total tuition expenses are: 25 + 30 + 35 + 35 + 40 = 165

What I need to do is come with a dataset listing of each person's
PERSONAL expenses after the accounting for the scholarship money. For
example, the COST column for DAVE represents the amount he must
personally come up with to cover the cost of tuition. It is safe to
assume that a person will first choose to exhaust his/her scholarship
money before dipping into their own wallet.

Dataset PRSN_EXP:
PERSON  YEAR    COST
DAVE    2004    0  --> tuition is completely covered by scholarship
DAVE    2005    0  --> tuition is completely covered by scholarship
DAVE    2006    10 --> tuition cost for this year exceeds the remaining
scholarship money balance
DAVE    2007    35 --> no dedection b/c scholarship money is completely
exhausted
DAVE    2008    40 --> all scholarship has been exhausted already
JOHN    2004    0
JOHN    2005    0
JOHN    2006    0 --> John graduates with a remaining balance in his
scholarship fund
KIM     2004    0
KIM     2005    0
KIM     2006    0
KIM     2007    0
KIM     2008    20

I am not sure how to handle the processing of deriving this PRSN_EXP
dataset from the first two datasets.

I would appreciate any suggestions/advice/pseudo code/articles with
similar ideas.

Thanks in advance for your help.