Hi.  Try this and see if it helps ...   * some data; data x; input x :$1. @@; datalines; 1 2 1 2 ;   * location of format library; libname mylib 'z:\';   * create a format ; * creates a CATALOG named FORMATS (default if no other name specified) in the library MYLIB; * NO NEED for that CNTLOUT option (useless); proc format library=mylib; value $noyes "1" = "Yes" "2" = "No"; run;   * extra step for now ... check to see if the new format is in the catalog within the library; proc catalog cat=mylib.formats; contents; run;   * tell SAS to include the new format catalog when looking for a formats; * since you did not give the catalog a name, SAS infers that it chould look for a catalog named FORMATS in MYLIB; options fmtsearch=(mylib);   * use the format (do not need STRIP function ... useless); data xplus; set x; y = put(x,$noyes.); run;   output from PROC CATALOG ... Contents of Catalog MYLIB.FORMATS # Name  Type     Create Date        Modified Date           Description 1 NOYES FORMATC 09/15/2015 11:46:53 09/15/2015 11:46:53   data set XPLUS ... Obs x y 1   1 Yes 2   2 No 3   1 Yes 4   2 No   More info ... Yes, We Can... Save SAS® Formats http://support.sas.com/resources/papers/proceedings12/048-2012.pdf   ... View more

Hi.  It might help to read a classic paper by Bob Virgile (SAS guy "extraordinaire") ...   How MERGE Really Works http://www.lexjansen.com/nesug/nesug99/ad/ad155.pdf ... View more

Hi, this works for me ...   proc printto log='z:\log.txt' print='z:\print.txt'; run;   proc means data=sashelp.class noprint; run;   proc printto; run;   There's nothing in the output window, here's the LOG ...   390 proc printto log='z:\log.txt' print='z:\print.txt'; 391 run; NOTE: PROCEDURE PRINTTO used (Total process time): real time 0.00 seconds cpu time 0.00 seconds ... View more

Hi, since you have the ARRAY statement after the SET statement, you should not have to explicitly state that the array EXPERIMENT is CHARACTER (it's implicitly known at that point given that the STUDY variables are character).   One thing that might be causing a problem is this line ...   if experiment{i} in ('106','112','123','133','135','139'')then type{i}=1;   You have an extra QUOTATION MARK at the end of the value list in the IN clause.     Also, this should cause a comment in the LOG about variable type conversion ...   else if experiment{i}=. then type{i}=.;   EXPERIMENT is a CHARACTER array, so that period for a missing value should be a space (" ") or just use ...   else if missing(experiment{i}) then type{i}=.;   and you don't have to worry about variable type.   A suggestion, how about using a FORMAT to recode you varaible values ...   proc format; value $e2t '106','112','123','133','135','139' = '1' '103 - PartA','103 - PartB', '109 - PartA','109 - PartB', '126 - PartA','126 - PartB' = '2' 'Ethics train', ' ' = ' ' other = '0' ; run;   data have; infile datalines dsd; input study1 :$15. @@; study2 = study1; datalines; '106', '112', '123', '133', ,'135','139' '103 - PartA', '103 - PartB', '109 - PartA', '109 - PartB' '126 - PartA', '126 - PartB' 'Ethics train', ' ', '999999999' ;   data want (drop=j); set have; array experiment(2) study: ; array topic(2); do j=1 to 2; * use the INPUT function to make TOPIC a numeric variable; topic(j) = input(put(experiment(j),$e2t.),1.); end; run;   If you print data set WANT you'll see that all the recoding works with a format. ... View more

Hi, here's another approach that uses LEXPERM to generate the permutations of nodes (same answer as the other posted solution with some paths shown in the reverse order) ...   * generate permutations of four nodes; data perm; array x(4) $2 ('32' '14' '55' '91'); do j=1 to fact(4);    rc=lexperm(j, of x(*));    output; end; keep x:; run;   * identify duplicates (left-2-right = right-2-left, output only one of the pair); data check; retain x0 x5 '11'; set perm nobs=howmany; s1 = catt(of x1-x4); do j=_n_ to howmany;    set perm point=j;    s2 = catt(of x4-x1);    if s1 eq s2 then leave; end; if s1 eq s2 then output; keep x:; run;   data set CHECK ... Obs x0 x1 x2 x3 x4 x5  1  11 91 55 32 14 11  2  11 55 91 32 14 11  3  11 91 32 55 14 11  4  11 32 91 55 14 11  5  11 55 32 91 14 11  6  11 32 55 91 14 11  7  11 91 55 14 32 11  8  11 55 91 14 32 11  9  11 91 14 55 32 11 10  11 55 14 91 32 11 11  11 91 32 14 55 11 12  11 91 14 32 55 11     ... View more

Hi, sugestion for renaming the variables in my previous post (insert just after PROC TRANSPOSE)   data y; set y; if length(instr) eq 4 then instr = catt('ins0',char(instr,4)); run;   That would convert INS1 to INS01, INS2 to INS02, etc. and the table columns would be in variable number order. ... View more

Hi.  If you don't mind creating another data set, you could transpose your data prior to PROC FREQ ...   * create some data, 100 observations, 14 variables (an ID plus 13 instructions) data x; array ins(13); do id=1 to 100; do j=1 to 13; ins(j) = int(200*ranuni(1234)); end; output; end; keep i: ; run;   * create a new data set with only two variables, INSTR (instructions 1 to 13) and COL1 (value of instructions 1 to 13); proc transpose data=x out=y (drop=id) name=instr ; var ins1-ins13; by id; run;   * use PROC FREQ on the new data set; proc freq data=y; table col1*instr / norow nocol nopercent; run;   Hopefully you won't mind that instructions 1 and then 10 through 13 appear in the first 5 columns.  If it's an issue, you could always change the values of variable INSTR in data set Y (make INSTR1 -> INSTR01, INSTR2 -> INSTR02, etc.). ... View more

CHROME cleanup of CACHE and BROWSING HISTORY is of NO HELP in getting rid of the 404 problem   no issues in IE 11     ... View more

Hi, you could try ...   data x; input name :$10 @@; datalines; KAREEM MARK LIAM MICHAEL ;   data last_m; set x; if left(reverse(name)) =: "M"; run;   data set LAST_M Obs name 1 KAREEM 2 LIAM   ... View more

Hi, here's a data step solution ...   data want; want = 0; do until (last.id);    set have (in=one) have;    by id;    if one and have then want = 1;    if ^one then output; end; run;   It's based on Howard Schreier's idea of a "self-interleave" ...   Interleaving a Dataset with Itself: How and Why http://www.lexjansen.com/nesug/nesug03/cc/cc002.pdf ... View more

Hi ... you can try this ...   * create a data file (no data set created); data _null_; * specify the name of the data file; file 'c:\new.txt'; * create some numeric variables and give them all a value of 99; retain x1 x2 x3 99; * write the values of the variables to the file specified in the FILE statement; put x1-x3; run;   * read the data file you just created and create a data set; data new; * specify the location of the data file; infile 'c:\new.txt'; * read the data; input x1-x3; run;   * look at your data set; proc print data=new; run;   Another way to remember the difference between FILE and INFILE is to think of the statements PUT and INPUT.  PUT statements are associated with FILE (write to a file) and INPUT statements are associated with INFILE (read data from an already existing file). ... View more