I greatly appreciate the help. I'm still new to SAS - but I've started the Programming Essentials I course, so hopefully that along with asking questions will help me learn. Thank you again! ... View more

I think the idea is to track the person overtime (again, I was given this to "explore"). So a dataset like so: 001     09/FA     other fields/variables 001     10/FA     other fields/variables     001     11/FA     other fields/variables would become: 001 09/FA     (09/FA related variables)     10/FA (related variables)     11/FA (related variables) I don't foresee having to deal with century boundaries at this time, 10 years seems to be the most we would deal with at any one time. It's not that I want to split into 6 years, I want to split into n years, where n is set by the user in a macrovariable (num_years) at the beginning of the program (along with the year_cohort). ... View more

Okay, I was given the following with the idea of only changing two variables - the starting year (year_cohort) and the number of years.  The comments below are the comments as I was given them. %let year_cohort=03;/*the fall semester belongs to academic year + 1*/ %let AY_start = &year_cohort + 1;/*The fall academic year*/ %let AY_end  =  &year_cohort + 6; %let AY_YEAR1=04; %let AY_YEAR2=05; %let AY_YEAR3=06; %let AY_YEAR4=07; %let AY_YEAR5=08; %let AY_YEAR6=09; /*EACH GRAD DATA SET REPRESENT AN ACADEMIC YEAR OF GRADUATION DATA. GRAD_1, FOR EXAMPLE, REPRESENTS GRADUATION DATA */ /*STARTING FROM C0HORT FALL SEMESTER (DECEMBER GRADUATION) THROUGH THE SUMMER GRADUATION ENDING THE ACADEMIC YEAR.*/ DATA GRAD_1(KEEP = CRED_ID Id College_1 MAJOR_1 CIP_1 Degree1_C1 Degree_1 Lvl_1 SEMESTER_GRADUATED_1 GRADUATE_DATE_1 ) GRAD_2 (KEEP = CRED_ID ID College_2 MAJOR_2 CIP_2 Degree1_C2 Degree_2 Lvl_2 SEMESTER_GRADUATED_2 GRADUATE_DATE_2 ) GRAD_3 (KEEP=  CRED_ID ID College_3 MAJOR_3 CIP_3 Degree1_C3 Degree_3 Lvl_3 SEMESTER_GRADUATED_3 GRADUATE_DATE_3 ) GRAD_4 (KEEP=  CRED_ID ID College_4 MAJOR_4 CIP_4 Degree1_C4 Degree_4 Lvl_4 SEMESTER_GRADUATED_4 GRADUATE_DATE_4 ) GRAD_5 (KEEP=  CRED_ID ID College_5 MAJOR_5 CIP_5 Degree1_C5 Degree_5 Lvl_5 SEMESTER_GRADUATED_5 GRADUATE_DATE_5 ) GRAD_6 (KEEP=  CRED_ID ID College_6 MAJOR_6 CIP_6 Degree1_C6 Degree_6 Lvl_6 SEMESTER_GRADUATED_6 GRADUATE_DATE_6) ; merge GRADUATE (in=a) cip; by major LVL_DEGREE; if a; LENGTH SEMESTER_GRADUATED $3.; GRADUATE_DATE=datepart(GRAD_DATE);/*ORACLE DATES HAVE TIME COMPONENTS TO THEM. THIS SAS FUNCTION JUST TAKES THE DATE PART.*/ /*CREATE SEMESTER GRADUATED VARIABLE:First two characters contain the ACADEMIC year. THREE GRADUATION MONTHS IN AN ACADEMIC YEAR: 12, 05, 08*/ IF  PUT(MONTH(GRADUATE_DATE),z2.)='12'  THEN SEMESTER_GRADUATED=SUBSTR(ACADEMIC_YEAR,1,2)||'1'; ELSE IF PUT(MONTH(GRADUATE_DATE),z2.)='05' THEN SEMESTER_GRADUATED=SUBSTR(ACADEMIC_YEAR,1,2)||'2'; ELSE IF PUT(MONTH(GRADUATE_DATE),z2.)='08' THEN SEMESTER_GRADUATED=SUBSTR(ACADEMIC_YEAR,1,2)||'4'; *keep  Id College MAJOR CIP Degree1_w Degree2 Lvl SEMESTER_GRADUATED GRADUATE_DATE; /*An array of character values and not variable names. Hence, the $ and parenthesis around the list*/ array Grad_year {1:6} $ ("&AY_YEAR1" "&AY_YEAR2" "&AY_YEAR3" "&AY_YEAR4" "&AY_YEAR5" "&AY_YEAR6"); /*CREATE VARIABLES FOR EACH GRADUATION ACADEMIC YEAR*/     %MACRO CREATE_VAR(START,END); %DO I=&START %TO &END; IF ACADEMIC_YEAR=grad_year{&I} THEN do; /*COMES FROM CIP FILE*/ College_&I= College_C; CIP_&I=CIP; Degree1_C&I=DEGREE1_C;Degree_&I=DEGREE; Lvl_&I=LVL_C; /*COMES FROM GRADUATE FILE*/ MAJOR_&I=major;  SEMESTER_GRADUATED_&I=SEMESTER_GRADUATED; GRADUATE_DATE_&I = GRADUATE_DATE; FORMAT GRADUATE_DATE_&I MMDDYY8.; OUTPUT GRAD_&I; end; %END; %MEND CREATE_VAR; %CREATE_VAR(1,6); RUN; ----------------- So you can see the questions I have above, and where I am trying to go with them. Sorry about not providing this at the beginning. I wasn't sure sharing this much would be helpful. Again, the idea is to only have to change year_cohort= and to add a num_years variable to set the number of years, and have everything else fall into place from there. ... View more

You're right (of course) - I want macro variable AY_Year1 to have a value of 04 (in this case). So in the above table that I inserted, Grad_year1 is almost correct (I want it to say 04, not 4), but Grad_year2 is incorrect (-9, it should say 05). I suspect what I'm doing is just approached the wrong way to begin with. I've been using SAS for less than 6 months, and I haven't had much time to actually learn what I am doing, just "Googling" this and that, etc. This:      array Grad_year {&num_years} (&ay_start-&ay_end); does the same thing as this:      array Grad_year {1:&num_years} (&ay_start-&ay_end); which is to say, Grad_year1 = 4 (instead of 04) and Grad_year2 = -9 (instead of 05) %put &ay_year1 = 04 ... View more

Thank you, but it isn't working as expected. Here is my input: %let year_cohort=03; %let num_years=6; data _null_;     start=&year_cohort;     call symput('ay_start', put(start+1, z2.));     call symput('ay_end', put(start+&num_years, z2.));     do i=1 to &num_years;         call symput("ay_year"||compress(put(i, 8.)), put(start+i, z2.));     end; run; data test;     array Grad_year {1:&num_years} (&ay_start-&ay_end); run; and here is the log: 391  %let year_cohort=03; 392  %let num_years=6; 393 394  data _null_; 395      start=&year_cohort; 396      call symput('ay_start', put(start+1, z2.)); 397      call symput('ay_end', put(start+&num_years, z2.)); 398 399      do i=1 to &num_years; 400          call symput("ay_year"||compress(put(i, 8.)), put(start+i, z2.)); 401      end; 402  run; NOTE: DATA statement used (Total process time):       real time           0.00 seconds       cpu time            0.00 seconds 403 404  data test; 405      array Grad_year {1:&num_years} (&ay_start-&ay_end); WARNING: Partial value initialization of the array Grad_year. 406  run; NOTE: The data set WORK.TEST has 1 observations and 6 variables. NOTE: DATA statement used (Total process time):       real time           0.00 seconds       cpu time            0.00 seconds And the dataset is as follows: Grad_year1 Grad_year2 Grad_year3 Grad_year4 Grad_year5 Grad_year6 4 -9 . . . . It's reconciling to 4 -9 instead of 04-09 (in this case). I'll keep trying things along those lines. Thank you again! ETA: I've also tried using spaces ( &ay_start - &ay_end ) and quotes ("&ay_start" - "&ay_end") with different, although still not working, results. ... View more

Additional question (should this be asked separately?) Can I create one macro variable that is essentially the concatenation of those created in the do loop? I want to replace this: array Grad_year {1:6} $ ("&AY_YEAR1" "&AY_YEAR2" "&AY_YEAR3" "&AY_YEAR4" "&AY_YEAR5" "&AY_YEAR6"); With something like this: array Grad_year {1:&num_years} $ (&ay_years); ... View more

What does the year_count = 1; in the middle of the data _null_ step do? ... View more

Thank you both! The first I tried worked as intended, though I may need to explore the second for another piece of this puzzle at a later date. Thank you again! ... View more

Arthur, Thank you so much for your additional methods.  It may take me a while to fully unpack and adapt your method, but I definitely love learning different approaches. I am using year as a string because I am using it as the vbar in a gchart. When I tried this with numerical year, it would create classes (2002.5 2007.5, etc), and I need things reported by instances per year. There may be another way, but I was handed this as a kind of last minute thing, so much of what I did was "brute force" and not particularly elegant, even by the standards of what little knowledge I have. I have completed the report, which used nearly 400,000 records split across 11 years and was reported differently on 27 pages, each with a gchart and a table that reported percentage growth of different variables per year.  Now that it is done, I plan on exploring various method and seeing if I can tighten it up a bit.  I've used macros in the past before, but due to the rushed nature of this, and how I (at least initially) felt that each page was different enough to make the macro not my best option. The lag function, of which I was not previously aware, made a huge difference - but I also like elegant solutions when I understand them. Thank you all again! ... View more

Thank you both so much! I ended up using a little from both - the key was the lag function!  Once I was aware of that, it was MUCH easier. In my original "code sample", I included the ( * 100 ) bit, but I took that out and opted for the PERCENT8.2 format instead. Thank you again! ... View more