Leapyears are every four years EXCEPT for century years which are only leapyears if they are divisible by 400. So, 2012 is a leap year (divisible by 4), 1900 was not a leapyear (it was a century year but was not divisible by 400), 2000 was a leapyear (century year divisible by 400). ... View more

It doesn't look to me like it is sorting at all. Even if Date_add was a character variable (which it shouldn't be) you wouldn't get those results. Are you getting any error messages? Are you looking at all_patients immediately after that step finishes? ... View more

As Tom and art297 said, it's not clear what you want. Your first sort will show (for example) all records with a date_add of 01JAN2011 then all records (for axample) with a date_add of 02JAN2011, and so on. For any given date it will show all patients in order that have that particular date. For any given date and patient it will show all times in order, and so on. One possibility comes to mind - what are the types (character or numeric) of those fields? The type won't affect whether you can sort, but may affect the resulting order (e.g. "10AUG1999", "10DEC1999", "10FEB1999", "11AUG1995" are in ascending order if they are in a character field). ... View more

If you want an illustration of what Doc@Duke is getting at, right-click on one of your .CSV files in Windows, choose Open With, and open it in Notepad instead of Excel. You will see that there is NO formatting information; it just contains comma-separated data. ... View more

Remember that what a macro statement generates need not be a complete SAS statement. So, in the case of your %ELSE you need to do two completely distinct things - finish the %ELSE statement and finish the SET statement. With only a single semicolon SAS can't tell that your SAS statment is finished because the semicolon has already been used to finish the %ELSE. ... View more

That question can be interpreted more than one way. The others have mentioned the various forms of an INPUT statement, which may very well be what was intended. However, looking at the question from a different point of view, four methods to read a file that I can think of would be INFILE, PROC IMPORT, LIBNAME (using a non-default engine) and %INCLUDE. ... View more

mick_g - your problem is that you haven't set the eof variable anywhere, so your if not eof then do... block never executes. You need to add end=eof to your set statement ... View more

> Then, I don't understand what would be the expected > result. One responder indicated a feature like > Excel's sumproduct function which also only provides > the sum of only non-missing products as well. What > value would you expect to result if there was a > missing value? In the original question he was asking for something similar to SUM but for products. SUM(1, 2, . , 4, . ) results in 7 even though 1+2+.+4+. results in . I understood his question as wanting something like PRODUCT(1, 2, ., 4, .) giving a result of 8 even though 1*2*.*4*. results in . . ... View more

That wouldn't solve the problem. The SUM function takes the sum of the non-missing arguments. What is being asked for is a function that would take the product of the non-missing arguments. Saying something like SUM(A*B) won't help; the result would be missing if either A or B was missing. ... View more

That is just going to rename them, not convert them. ... View more

Firstly, you should probably make your DrinksWeek_Last_3_Months a numeric variable rather than a character variable (that's what the "Character values have been converted..." alludes to). From that code segment, it looks like all incoming records have a value of " " or "99" for DrinksWeek_Last_3_Months. If that is not the case, then you would need to post a sample of your data. ... View more

the first put (that you are using to convert to a numeric) should be input. ... View more

I think you just want the DBMAX_TEXT=n option on your libname statment or dataset option. Check out the documentation on it. The value you set becomes the maximum width that will be read. The maximum is 32767. HOWEVER, one thing I did notice is that my programs did get slower the higher I set it, so be careful about going overboard with it. (I don't have any examples because I switched jobs recently and no longer have a need for it). Message was edited by: Douglas Martin ... View more

If you're doing stats based just on the idno variable then you don't need to combine anything - just run proc means on the file2 dataset. So, lets assume you want to display the name (presumably from file1) in place of idno, and do stats on result (presumably in file2). I'm doing this off the top of my head, so there may be typos. I've included some other stats besides mean just to give you a better example. libname cd "D:\"; data combined; merge cd.file1 cd.file2; by IDNo; run; proc means data=combined mean std median min max; class name; var result; run; If that's not enough explanation, you really need to consult the documentation. ... View more

Do you mean something more than doing a merge or SQL join? data combined; merge cd.file1 cd.file2; by IDNo; run; If you like SQL you can do the equivalent in PROC SQL. ... View more