If your program is a linear process that's meant to be run from top to bottom in its entirety (which I require everyone on my team to do) as opposed to highlighting and running chunks, then having the input and output dataset names the same is much more space-efficient, not to mention less likely to introduce errors (setting the wrong dataset). We do everything in Linux and launch everything from the command line. Your mileage may vary. ... View more

I think the difficulty with this question is that you've actually created an over-simplified example of what you actually need.  Here, you've got a formula table that you're joining (in a Cartesian way) to your data, but the formula table only, in this example, contains a single formula, so it's not clear how the join would be modified if it contained multiple formulas that would be used conditionally based on, I guess, the data itself.  It would be easier for people to respond if your example formula table contained at least 2 different formulas and that the data step that followed showed us how you would use those formulas.     In any case, I'm guessing you have a programming background that is maybe something other than SAS - nothing wrong with that - just saying that the way you're approaching this is very unusual and probably much more complicated than it needs to be.     With your current example, you could just skip the formula table (and the join) entirely and just run a macro, e.g.: %macro xy; x * y - 1 %mend; data want; set have; calc=%xy; run; ...but again, if you provide a more realistic example, we might be able to give you a better answer. ... View more

dim is a function - you have to use actual parentheses with it. ... View more

actually, the syscc thing doesn't seem to work. You might try the `abort abend` method in the edited version above. ... View more

OK, right, that isn't going to cause a system error.  If you wanted to do that, you could do something like: if product=' ' then do; put 'ERROR: ' plan=; abort abend; end; ... View more

And you've specified OPTIONS ERRORABEND? ... View more

That's very strange - you're saying that the log for 'driver.sas' only contains stuff from the 3rd program? Maybe others can explain that. That's definitely not the behavior I get when doing this kind of thing in Linux. ... View more

It's really just something as simple as creating a driver program (we'll call it driver.sas) and running that program: options errorabend; %include "program #1"; %include "program #2"; %include "program #3"; * ... ;     ... View more

There are some things wrong with your code - for one thing, you need a BY statement in your data step if you're going to use things like first.<variable> or last.<variable> -- in this case:  BY descending NUMREK; Secondly, the way you're defining C is just going to give you either a 1 or 0 (first.numrek is true or first.numrek is false).     I am not totally sure what you're trying to do, but here's how I would suggest you modify your DATA step (in this case, you don't actually need the BY statement):   data have; set have; retain C; if C=. then C=NUMREK; * populate C with first non-missing NUMREK (i.e., largest available), then not again ; NUMREK=coalesce(NUMREK, C+1); run; That does what you *asked* for, but I suspect that's not really what you want.  For one thing, by ordering as DESCENDING, the values of NUMREK are getting smaller and smaller as you go through the data, so I'm not really sure what you mean by "the maximum of the variable + 1".  The maximum value of NUMREK is the first non-missing value in your sorted dataset.   I imagine what you actually want is something like this: proc sort data=have; by numrek; run; data have; set have; retain C; if numrek > . then c=numrek; else numrek=c+1; run; This populates a missing numrek with the largest value seen so far + 1.  ... View more

Only thing I can think is to try the PSMATCH with PCDK46 as the only explanatory variable in the PSMODEL statement and seeing if the associated SMD (from the StdDiff table output) is the same as what you're currently getting. I also don't know offhand what the 'nlargestwgt' argument is doing and whether that's affecting things. ... View more

Should work just like this: ** THEN, IN YOUR PROGRAM #1 PUT THIS AT THE **END** ; %startstop(beginning=0, lib=<YOUR LIBREF>); ** and in PROGRAMS #2 and higher, put this at the beginning: ; %startstop(beginning=1, lib=<YOUR LIBREF>); ** ... and this at the end ; %startstop(beginning=0, lib=<YOUR LIBREF>); beginning=1 is just telling the macro to run the part of the code that's needed for the START of a program (for all those #2 and higher) -- it checks whether the 'continue' file exists, aborts if not, and deletes it and continues if so.   beginning=0 tells it to run the part of the code needed for the END of each program (creates the 'continue' file).     That should work, but I think the easiest solution is just the driver file proposed by @SASKiwi  though -- just a program that %includes all of your other programs, in order.  If you do that, I would suggest adding:  OPTIONS ERRORABEND; to the top of the driver program.  This will cause the program to *stop* if it encounters an error rather than going into syntax check mode and continuing.  If you don't add this option and the programs have never run before (i.e., there are no datasets in existence from a previous run), syntax-check mode still seems to create empty datasets, which can be a little weird / misleading.   ... View more

The variable PCDK46 has exactly 2 levels and no missing values? ... View more

Wow, that POLYLINE thing on p. 16 is great - I've never seen that before. I've made those kinds of brackets connecting different parts of a plot by building the shape with several vector statements, but it's incredibly awkward that way - this is perfect! ... View more

You could MAYBE do 3 separate YAXISTABLE statements (one for each column) and position them next to each other. As for italicizing a single character - I really doubt that could be done with SGPLOT. ... View more