For the first observation with k=2, did you mean for the result to be 100 (continued from k=1)? It makes a difference in the list of possible solutions. ... View more

SAS has published a way to do what  you ask:   https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/mcrolref/p0j1htu10wsx9tn1mig5g0b8mxxb.htm     ... View more

Changing TIME to numeric: data new_ds1; set ds1; numtime = input(substr(time, 6), 3.); run; SUBSTR will start with the 6th character, and capture all remaining characters.  INPUT converts character to numeric (here, and also in most cases). ... View more

Just a few items to consider: Why do this at all?  The data appears to be useful in its current form. Why keep TIME as a character variable?  It would be more useful as numeric, and would then easily sort in the right order. When there is no measurement (such as time_15 in DS1), why change missing values to a zero?  Zero is a perfectly good measurement on other observations, so why use that to represent a missing value? ... View more

I hear you.  But SAS disagrees.  The message it is giving means that at least two of your incoming data sets contain multiple observations for a USUBJID.   Here's a program you can run to find the source of the problem in the DM1 data set: proc freq data=dm1 noprint; tables usubjid / out=dm1_multiples (where=(count > 1) drop=percent); run; proc print data=dm1_multiples; run; Unfortunately, you will need to test all the data sets.  Since you are expecting no such problem, it is possible the problem lies with missing values:  multiple observations where USUBJID has a missing value.  ... View more

The best solution would likely be to fix the data sets before you merge them.  More specifically, why should the same USUBJID have two observations in the same data set?  In fact, it has two (or more) observations in more than one incoming data set.  That's what the message tells you.    If you can't make USUBJID unique, then you have to choose among more complex scenarios  For example, what result do you want if the same USUBJID appears three times in one data set, and twice in another?  That's a choice that depends on your knowledge of what is in the data (or what should be in the data).  Start with making USUBJID unique if that's at all possible. ... View more

@Tom , you're absolutely right.  Good point.   After sorting, the "decluttered" version of my solution would be: data want; do i=1 to 5, 5 to 1 by -1; set have; output; end; run; If HAVE runs out of observations before a loop completes, that would be a normal end to the DATA step. ... View more

Logically, I think this is the right way to get what you are describing.  After sorting:   data want; do until (5=4); do i=1 to 5, 5 to 1 by -1; set have; output; end; end; run; It might seem strange to use 5=4 as the UNTIL condition.  That just means the DATA step keeps on going until the SET statement ends it by running out of observations to read.   There are other legitimate questions to ask.  Is this the best statistical way to divide up your data?  (Probably not.)  Why didn't your earlier attempt with two SET statements work?  (Each SET statement operates independently of the other.  Both start reading the first observation from the incoming data, so I expect you have 649 + 645 observations from that attempt.)  Those questions can be addressed in more detail.  But for now, the above program should get you what you asked for. ... View more

@whymath ,   This particular trick?  No.  But even beginning programmers use logical expressions all the time.  For example, consider: if first.state then do; There is no comparison there.  First.state is either 1 or 0, and the software considers 1 to be true and 0 to be false.  So there is no need for: if first.state=1 then do; Actually, the evaluation of logical expressions goes beyond 1 and 0.  The software considers 0 and missing values to be false, and all other values (including negative numbers) to be true.  So consider a simple statement: a = b / c; When c is 0, the software notes that division by zero took place, and tracks how many times it happened.   It also runs up the bill, taking vastly more CPU time.  Similarly, when c is missing, the result is that a missing value gets generated.  Again, the software runs up the bill (CPU time, that is), with no usable numeric result.  So if you have lots of missing values of zeros for C, it can be faster to use this statement: if c then a = b / c; The IF clause treats C as a logical expression (false for missings and zero, but true for all other values.  So there is no attempt to calculate anything when C is missing or zero.  I don't think I can come up with a realistic use for: do i=1 to 5, j=5 to 7; ... View more

Too many pieces to this that I don't understand.  But the one I do understand is this.    Macro language %IF %THEN statements do not execute for every observation in a SAS data set.  They merely generate DATA step code.  So this statement can't work as intended: %IF missing(&var._min_day) %THEN %DO; Even adding %SYSFUNC around the MISSING function won't help.  You would need to convert to DATA step IF THEN statements instead. ... View more

One more (untested) variation came to mind:   druguse2 = ^^ marijuanause2 + 2 * ^^illegal_druguse2; ... View more