before getting distracted by the roman numerals, I intended to make use of the SORT FUNCTION since I don't get to use it very much.  Anyway, following my own "Juolas" Metatheory I did not want to assume the categories would be alphabetical.  BUT maybe we can, and we can use SORT.: data have; infile datalines dlm=':'; input ID Categories :$100.; datalines; 1:(i)cyt,(ii)end therapy,(iii)targ 2:(ii)cyt,(iii)targ,(ii)end therapy 3:(iii)end therapy,(i)cyt,(iii)targ 4:iii)targ,(i)cyt,(ii)end therapy 5:(iii)end therapy,(i)cyt,(iii)targ 6:iii)targ,(i)cyt,(ii)end therapy 7:(i)cyt,(ii)end therapy,(iii)targ 8:(ii)cyt,(iii)targ,(ii)end therapy 9:(ii)cyt 10:(iii)targ 11:(ii)end therapy ; data want1; set have; length cat1-cat3 $20; array cat[3] $ cat1-cat3; do i = 1 to 3;/*assuming 3 categories*/ cat[i] = strip(scan(scan(categories,i,","),2,")")); end; _rc=sort(of cat1-cat3); category=catx(", ",of cat1-cat3); keep id category; run; ... View more

I stand corrected ... View more

Do you have any other data in the two data sets in common such as date of birth, residence (city, state/province, postal code), telephone number or such?   If you have additional information such as above you may want to investigate Link Plus from the CDC. Website: https://www.cdc.gov/cancer/npcr/tools/registryplus/lp.htm   This is a probabilistic matching program that will use multiple pieces of information and provide a probability of a match between two records. It will report the probability that two records match so you might be able to reduce manual coding a bit and only spend time with the problem cases. I've never used it this tool with only names but have used it to identify when a data base was migrated and determined individuals whose names were spelled differently or had changes in date of birth, race, ethnicity or sex when the data was transferred between data systems. One thing is that the software is free. It does want text files with column headers (SAS makes those pretty easy with Proc Export) and you provide the rules for which fields to compare between two files. Example one file might have date of birth information in a column DOB and the other BirthDate. So you tell the program those columns are to be compared.   If you have tens of thousands of cases and only names I would call this a "job security" problem because there is so much garbage you are going to have to deal with. You haven't even addresses things like Robert, Bob and Bobby or similar nickname patterns.   One problem I have with eliminating practically any give word like "Old" from these fields are some of the actual names. Consider celebrity children names like Moon Unit Zappa, or anglicized Native American names. There are also parents out there that go a long way to get interesting names for children, Though matches on those name may be more obvious it might also mean you don't want to just throw out words.   A proactive approach would be to go the source of the data and say they have to watch their data entry standards a bit more closely.   I say this from dealing with things like Address data for geocoding that had addresses with text like "see the woman in the building behind the main house"   You may also want to make sure you aren't dealing with multiple character sets such as accented letters before attempting this. ... View more

Sounds like you don't want to MERGE that data at all.  Use a SET statement with the INDSNAME= option to indicate which dataset contributed the observation. You could then store the name into a permanent variable and use that in conjunction with PROC TRANSPOSE to generate new variable names. data step1; set ds1 ds2 ds3 indsname=dsn; by id1 id2; dsname = upcase(scan(dsn,-1,'.')); run; proc transpose data=step1 name=variable out=step2; by id1 id2 dsname notsorted; var dob name; run; proc transpose data=step2 out=want(drop=_name_) delim=_; by id1 id2 ; id variable dsname; var col1; run; ... View more

Note that when using the FILEVAR= option it is better to use DO WHILE() instead of DO UNTIL() to avoid any empty file in the list causing the data step to stop early because it read past the end the input file. infile IN filevar=FILENAME end=EOF; do while (not EOF); input; if find(_INFILE_, TEXT, 't') then output; end;   ... View more

8.2 Update 4.  I'll edit the post.   I can use another computer, but I'm going to need to get back to you to arrange that. ... View more

The Cambridge Dictionary lists one of the uses of "of" as "relating to".  In this sense, the developers used the word well.  They used "of" to relate a function to a grouping of variables.    However, to @Tom's point, if we discuss this long enough, the Dictionary of Internet Slang's definition of "of" may soon apply to all of us. 😉   What does "of" mean in slang?   The Meaning of OF: OF means "Old Fart"      ... View more

NO ... View more

No idea.  but I want to use this macro.  I didn't see it when it came out.  Thanks for introducing it to me.   There's got to be answers for you on the Github page.  IDK any but I'll put the link here just to be a good forum participant:  GitHub - RhoInc/sas-sankeybarchart: A set of SAS macros for creating longitudinal bar charts with Sankey-style overlays.     ... View more