How about just using "data condition" and PUT to write some code that gets %INCed.

data _null_; set condition; if _n_ gt 1 then put 'Else ' @; put 'IF (' condition ') then ' result ';'; run;

You would want to write this to a file of course. I wish the OP would post a better sample of his actual data as he implies that it is more complicated.

On 9/10/09, Nallapeta1, Kiran (GE Money, consultant) <Kiran.Nallapeta1@ge.com> wrote: > Perry... I think this is what you are looking for. The below code builds > IF-ELSE statements dynamically from a table. > > ------------------ > /* Building Conditions and Results table */ > data condition; > length condition $50. result $50.; > infile cards dlm=',' missover pad; > input condition $ result $; > cards; > a < 3, z=0 > a > 3 and a lt 10, z=2 > a in (2 3 4), z=9 > a ne 0, z=1 > ; > run; > > /*Building IF-ELSE Logic/String*/ > %macro build_code; > proc sql noprint; > select condition, result > into :conds separated by "#", :res separated by "#" > from condition; > quit; > > %global fullstr; > > data condition2; > do i = 1 to &sqlobs; > cond = scan("&conds", i, "#"); > rest = scan("&res", i, "#"); > > str = "if " || cond || " then " || rest || ";"; > if i ne &sqlobs then str1 = str || " else"; > else str1 = str; > > length full_str $300.; > full_str = compbl(full_str || str1); > end; > call symput("fullstr", full_str); > run; > > /*Prints the IF-Else String to log*/ > %put &fullstr; > > /*Apply the IF-ELSE logic to below sample data*/ > data sample; > a=6; output; > a=3; output; > a=4; output; > a=50; output; > run; > data final; > set sample; > &fullstr; > run; > > %mend build_code; > > %build_code; > ------------------ > Thanks, > Kiran > > > -----Original Message----- > From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of > Perry > Sent: Thursday, September 03, 2009 6:05 PM > To: SAS-L@LISTSERV.UGA.EDU > Subject: Table Driven Waterfall > > I'm hoping that someone out there has done something similar to what I > need as I'm really looking for assistance in figuring out an ideal > table structure for my task. > > I have one dataset of selection criteria, basically a list of many > simple logical constructs, e.g. (x le 10000) or (y not in > ('A','B','C'). I want to define another table that combines (by > reference) any number of these simple criteria into a cascading set of > if..then..else statements. > > In other words, I want to be able to create macro arrays to generate > statements like: > > if criteria1 then .... ; else > if criteria2 then .... ; else > > etc. > > However, criteria<n> can (and will) be more complex than just a single > construct in the selection criteria table. > > Hopefully, I have provided enough information to convey what I am > looking for. Eventually, I could probably come up with a workable (but > probably flawed) solution, but am hoping to tap into the vast > expertise and experience of the list to see if someone has already > implemented something similar. > > TIA > Perry >