Show us, code, how you "merged" the two data sets.   And why are these named "export" if that has any bearing on the actual problem.   A general comment: having two values in a single variable such as that 33 (+12) almost always means extra work for very little gain over having two variables, such as Enrolled with a value of 33 and EnrollmentChange or similar name with a value of 12.   Please provide example data in the form of data step code.   This is what I would do if this were my project: data dataset_a; input Group $ Screened Enrolled Events; datalines; ABC 40 21 3 DEF 24 21 4 GHI 55 41 5 JKL 24 2 6 MNO 42 23 8 ; data dataset_b; input Group $ Screened Enrolled Events; datalines; ABC 89 33 13 DEF 56 43 4 GHI 88 45 22 JKL 67 23 34 MNO 68 23 23 ; /* sort both sets by Group variable if not in order*/ data want; merge dataset_a (rename=(enrolled=enr1 events=event1) ) dataset_b ; by group; enrollchange= enrolled-enr1; eventchange = events - event1; drop enr1 event1; run; Why? Because someone may want statistics on the changes such a maximim, minimum, mean or similar that can't be done with values that would have to be character such as 33 (+12). Just finding which of groups had the largest/smallest change with that character value would be an exercise in parsing the data back to a change value to compare them. ... View more

data work.old; length id instance 8 group term term_2 $20; input id group instance term term_2 ; infile datalines dlm="," dsd; datalines; 1, 1, 1, AAAA, one 1, 2 indirect, 2, CCCC, seven 1, 2 indirect, 3, DDDD, seven 2, 1, 1, ASFD, two 2, 2 indirect, 2, DDA, three 3, 2 direct, 1, ERQE, one 4, 1, 1, ABCD, seven 4, 1, 2, ABCD, five ; data work.new; length id instance 8 group term term_2 $20; input id group instance term term_2 ; infile datalines dlm="," dsd; datalines; 1, 1, 1, AAAA, one 1, 2 indirect, 2, CCCC, seven 1, 2 indirect, 3, DDDD, seven 1, 2 indirect, 4, DFDD, two 2, 1, 1, ASFD, two 2, 2 indirect, 2, DDA, three 3, 2 direct, 1, ERQE, one 4, 1, 1, ABCD, seven 4, 1, 2, ABCD, five 4, 2 direct, 3, CCCC, one 5, 1, 1, ERQE, seven ; data want; if _n_=1 then do; if 0 then set old; declare hash h(dataset:'old'); h.definekey(all:'yes'); h.definedone(); end; set new; flagT=1; if h.check()=0 then flagT=0; run; ... View more

It looks like your data sets:   Are already in order by ID Have no duplicates (more than one record with same ID)   If that's the case, the solution is simple:   data want; merge dataset1 dataset2 dataset3; by id; run; If you might have multiple records for the same ID, we need to know more about what the final result should look like. ... View more

You have to tell SAS where to LOOK for formats. Change the FMTSEARCH system option to include your new format catalog.  Since you didn't tell PROC FORMAT what name to use for the catalog it used FORMATS.  So you want to add either the libref TrialData to the search path or the catalog TrialData.formats. ... View more

You can get the final table by following proc sql few steps.  1. Get the cartesian Product of just Id and event. 2. Merge the step1 result with the original id_all table and get the common rows. 3. Get the subset from step1 which is excluded from step2 and add it back data event; input event $10.; datalines; Month1 Month2 Month3 run; data id_all; infile datalines missover; input ID event :$10. value value2; datalines; 5656 Month1 13 324 6565 Month2 31 465 4663 Month1 14 697 4242 1234 Month3 14 656 5421 Month1 12 647 run; proc sql; create table subset as select a.id, b.event from id_all a,event b; quit; proc sql; create table match as select a.id, b.event, a.value, a.value2 from id_all a left join subset b on a.id=b.id and a.event=b.event ; quit; proc sql; create table all as select * from match union (select id, event, . as value, . as value2 from subset except select id, event, . as value, . as value2 from match); quit; proc sort data=all; by id, event; run; proc print data=all; run;     ... View more

You only listed one name after the variables you specified, Ie N = N. You need to provide a variable name for each value, N = N1-N6 instead. Or don't specify anything and let SAS name it using the autoname option as illustrated by PaigeMiller. ... View more

data have; input ID $ VALUE $; infile cards truncover; cards; ABC 2 CDE EEA 2,3 AWE 4,5,6 DDE 5 QWE WQE 5,3,2 ; run; data want; set have; count=ifc(^ missing(VALUE),left(countw(VALUE,',')),''); run; ... View more

Use COUNT DISTINCT to count the number of unique values in a column.    FYI - I'm assuming you want the final code to be GROUP# + DIRECT, ie 3 Direct if 3 was the maximum. Your data doesn't show a 3 but if it was possible this still works.    EDIT: Modified to use calculated key word and only display results for the max group records.    proc sql; select *, max(group) as max_group, case when count(distinct group) > 1 and group = max_group then catt(calculated max_group, " Direct") else if group = calculated max_group then calculated max_group else " " end as new_group from data; group by id; quit; @pacman94 wrote: I have the following data ID Date group 1 1/2/2021 1 1 1/3/2021 2 2 1/2/2021 2 3 1/5/2021 2 4 1/4/2021 1 4 1/5/2021 2   What I would like to do is that for each ID, I need to take the max group and add additional groups if ID only has one instance then keep that new_group as is if the ID has more than one instance then new_group = '2 Indirect' like this   then only keep the new group ID Date group new_group 1 1/2/2021 1   1 1/3/2021 2 2 Indirect 2 1/2/2021 2 2 3 1/5/2021 2 2 4 1/4/2021 1   4 1/5/2021 2 2 Indirect   i did the following proc sql; select *, max(group) as group from data; group by id; quit;    ... View more

PROC FREQ can get you fairly close as well. Here's another way of doing it. Run the example code, as is and verify it meets your needs. If it does, then modify it for your data set. https://gist.github.com/statgeek/0c4aeec9053cf8050be18a03b842c1b9 Or consider trying this macro out: https://communities.sas.com/t5/SAS-Communities-Library/Demographic-Table-and-Subgroup-Summary-Macro-TABLEN/ta-p/634030 ... View more

And, in addition to @FreelanceReinhard's suggestion, keep in mind that your options such as headline, headskip and WIDTH= are ignored by ALL ODS destinations, these are LISTING only options and not relevant to ODS CSV. Cynthia ... View more

Hello. The original data looks like this: Group_Category Group Total Female Male Hispanic NotHispanic Group A GroupA_1 32 16 16 2 30 Group A GroupA_2 30 14 16 5 25 Group A GroupA_3 10 6 4 6 4 Group B GroupB_1 10 9 1 1 9 Group B GroupB_2 44 20 24 4 40 Group B GroupB_3 60 20 40 5 55 Female + Male = Total Hispanic + Not Hispanic = Total ... View more