Thanks guys - I've tried various versions of the code everyone has given me, but it's still not quite what I need. Perhaps it's because I'm not very advanced with SAS, but let me reiterate what I'm looking for because I must've not explained well. My data structure is like this    X . . . X X Y . Y Y . . Z . . Z Z    I need it to be like this  X  Y  Z  X  Y  Z  X  Y  Z      when I use this code:    data all.ap5; set all.stage9; array vars{900} VAR11_c1-VAR11_c900; obs = _n_; do i = 1 to 900; var11 = vars{i}; if not missing(var11) then output all.ap5; end; drop i var8_c1-var8_c1260 VAR11_c1-VAR11_c1260 VAR12_c1-VAR12_c1260 Participant__Nation_c1-Participant__Nation_c1260 var25_c1-var25_c1260 var26_c1-var26_c1260 VAR32_c1-VAR32_c1260 VAR34_c1-VAR34_c1260 VAR35_c1-VAR35_c1260 VAR36_c1-VAR36_c1260 VAR37_c1-VAR37_c1260 VAR50_c1-VAR50_c1260 VAR54_c1-VAR54_c1260 VAR55_c1-VAR55_c1260 VAR61_c1-VAR61_c1260; run; It gives me exactly what I need. However, instead of doing it for just one variable. I want to do it for the three variables at the same time, so I tried something like:    data all.test; set all.stage8; array vars{1440} VAR55_c1-VAR55_c1440; array vars1{1440} VAR61_c1-VAR61_c1440; obs = _n_; do i = 1 to 1440; var55 = vars{i}; if not missing(var55) then output all.test; end; do i = 1 to 1440; var61 = vars1{i}; if not missing(var61) then output all.test; end; drop i VAR55_c1-VAR55_c1440 VAR61_c1-VAR61_c1440; run; However, that does not work how I need it to..    If someone could point me in the right direction that would be great.    The variables I'm arraying are character variables with either a "Yes" or "No" as their values.    Thank you, R ... View more

Okay I tried.. data all.ap5; set all.stage9; array vars{900} VAR11_c1-VAR11_c900; array vars1{900} VAR25_c1-VAR25_c900; array vars2{900} VAR34_c1-VAR34_c900; obs = _n_; do i = 1 to 900; var11 = vars{i}; var25 = vars1{i}; var34 = vars2{i}; if nmiss(var11,var25,var34) = 0 thenoutput all.ap5; end; drop i var8_c1-var8_c1260 VAR11_c1-VAR11_c1260 VAR12_c1-VAR12_c1260 Participant__Nation_c1-Participant__Nation_c1260 var25_c1-var25_c1260 var26_c1-var26_c1260 VAR32_c1-VAR32_c1260 VAR34_c1-VAR34_c1260 VAR35_c1-VAR35_c1260 VAR36_c1-VAR36_c1260 VAR37_c1-VAR37_c1260 VAR50_c1-VAR50_c1260 VAR54_c1-VAR54_c1260 VAR55_c1-VAR55_c1260 VAR61_c1-VAR61_c1260; run; Unfortunately, that doesn't work. The three variables are character variables with a either a "Yes" or "No" in them (or blank, that's why I need to drop the missings). ... View more

Ballard,    data all.ap5; set all.stage9; array vars{900} VAR11_c1-VAR11_c900; array vars1{900} VAR25_c1-VAR25_c900; array vars2{900} VAR34_c1-VAR34_c900; obs = _n_; do i = 1 to 900; var11 = vars{i}; var25 = vars1{i}; var34 = vars2{i}; output all.ap5; end; drop i var8_c1-var8_c1260 VAR11_c1-VAR11_c1260 VAR12_c1-VAR12_c1260 Participant__Nation_c1-Participant__Nation_c1260 var25_c1-var25_c1260 var26_c1-var26_c1260 VAR32_c1-VAR32_c1260 VAR34_c1-VAR34_c1260 VAR35_c1-VAR35_c1260 VAR36_c1-VAR36_c1260 VAR37_c1-VAR37_c1260 VAR50_c1-VAR50_c1260 VAR54_c1-VAR54_c1260 VAR55_c1-VAR55_c1260 VAR61_c1-VAR61_c1260; run;         Thanks for this. However, the 'if not missing' part is critical to my code - is there a way to incorporate this in the code you offered? ... View more

Thank you ... View more

Thank you ... View more

Thanks - I know about the renaming, but is there a way I can merge the same way? There's gotta be a quicker way than typing out each variable name one by one. ... View more

Thank you! ... View more

Thank you. ... View more

PG - Finally got things figured out based on your longways code. Big help, this code is changing my life! Anyhow, thanks again.  ... View more

Haha, I follow your advice. Generally, I would like to keep wide format, for various reasons, but I dont mind doing certain computations long ways how you shown. Very helpful.        1 - So the merge code you just gave does not work well, since there is not an observation for every observation in the dataset (not every observation in the original dataset had an SIC that didnt match in the other array of SICs). Additionally, I need to have it matched up so that I know WHICH sic codes did not match, like so:    obs   nmisssics    dm1     dm2    dm3      ........etc. 1           2            4959     4215   4800     I suspect there is a line that could be added to the proc sql statement that created this table, but Im not sure what code to put in there:    /* Count the distinct missing sics in each observation */ proc sql; create table adtc.countSics as select obs, count(distinct sic) as nMissSics from adtc.missSics group by obs; quit;       2 - I tried to modify the code you gave me in order to create a table that counts how many matches there were and also list which codes were matched. If you wouldnt mind letting me know this looks:      /* Create two datasets. sics1 with psics 1-130 and portfolio with psics 131-520 */ data adtc.sics1 adtc.portfolio; set adtc.original; array sics1{130} psic1-psic130; array portfolio{390} psic131-psic520; obs = _n_; /* Unique ID for each observation */ do i = 1 to 130; sic = sics{i}; if not missing(sic) then output adtc.sics1; end; do i = 1 to 390; sic = portfolio{i}; if not missing(sic) then output adtc.portfolio; end; drop i psic1-psic520; run;   /* Extract the sics from Sics that appear on the same observation in Items */ proc sql; create table adtc.matchSics as select * from adtc.Sics1 as A where sic in (select sic from adtc.portfolio where obs=A.obs); quit;   /* Count the matched sics in each observation */ proc sql; create table adtc.countmatchSics as select obs, count(sic) as nMatches from adtc.matchsics group by obs; quit;     Again, I appreciate your patience! ... View more

PG - That worked great! I really appreciate your help with all of this. 2 more questions -    1. In that last table created by the last line of code you just posted (countsics), how do I also get it to include the distinct SICs as columns for each observations.    For example,    obs   nmisssics    dm1     dm2    dm3      ........etc. 1           2            4959     4215   4800     2.  What if I want to use similar code (similar to what you just sent me) to compare two arrays of numbers. The numbers are also currently arranged in columns (psic1-psic130 and match1-match30) However, this time I want it to count a comparison if there IS a match. For example, if the codes in the columns match1-match30 match ANY of the codes in psic1-psic130, I want it to create a table that counts how many matches there were and also list which codes were matched. Furthermore, there are repeats and missing values within psic1-psic390 and potentially missing values within match1-match30.    Thanks again for your invaluable help and patience, I dont take it lightly.        ... View more