This obviously depends on environment: 69 data _null_; 70 call sound(440,1000); 71 run; NOTE: Verwendet wurde: DATA statement - (Gesamtverarbeitungszeit): real time 0.00 seconds user cpu time 0.00 seconds system cpu time 0.00 seconds SAS On Demand, where sound makes no sense anyway. ... View more

Maxim 4: Try It.   But from experience I prefer option 2, either with the WHERE= dataset option or a WHERE statement. Unless you read from an external database, where you might want (depending on data structure) to use explicit passthrough first. ... View more

One of your PROC SORTs does not work because variable datadate is not in the dataset. You forgot it in the KEEP in line 49. ... View more

Given your examples, you want to round by 1000, not by 500. @Ronein wrote: Hello I want to round the number by 500 (up or down) for example  17524.5   will  be converted to  18000 25475.37  will  be converted to  25000 6475.3 will  be converted to  6000   what is the way to do it please?     ... View more

Define both z1 and z2 as group, calculate sums for groups defined by z1, and the compute the percentages: data have; input z1 $ z2 x w; cards; A 1 10 20 A 2 20 30 A 3 40 50 B 1 15 30 B 2 25 30 B 3 40 50 ; proc report data=have; column z1 z2 x pct_x w pct_w; define z1 / group; define z2 / group; define x / analysis; define pct_x / computed format=percent8.2; define w / analysis; define pct_w / computed format=percent8.2; compute before z1; x_sum = x.sum; w_sum = w.sum; endcomp; compute pct_x; pct_x = x.sum / x_sum; endcomp; compute pct_w; pct_w = w.sum / w_sum; endcomp; run; Result: z1 z2 x pct_x w pct_w A 1 10 14.29% 20 20.00% 2 20 28.57% 30 30.00% 3 40 57.14% 50 50.00% B 1 15 18.75% 30 27.27% 2 25 31.25% 30 27.27% 3 40 50.00% 50 45.45% ... View more

Why do you want only two icode columns? ... View more

Transpose to long, then sort: data long (rename=(dcdate=date)); set have; array ic {*} icode:; do i = 1 to dim(ic); if not missing (ic{i} then do; icode = ic{i}; output; enđ; end; keep id dcdate icode; run; proc sort data=long nodupkey; by id date icode; run; ... View more

Maxim 1: Read the Documentation INTNX Function: Alignment  ... View more

Transpose to a long dataset, so data (the years) is not hidden in structure. Maxim 19!! As soon as data is data, no dynamic code is needed, as the data drives the process. ... View more

Since no variables with names starting with nr_laks are present in your DATA step, and custid is automatically retained (because it is present in an incoming dataset), the RETAIN statement is completely irrelevant and does nothing. ... View more

Your SAS session needs to run with UTF-8 encoding to import such characters successfully. ... View more

This code can't work because don't is not a valid name for a macro. Only letters, digits and underscores can be part of a macro name. ... View more

The 49-169 points to a problem with unbalanced quotes or similar which has left the SAS system in an unstable state. Start a new session and run only your import code, then copy/paste the whole log into a window opened with this button: You can overwrite sensitive information with X's. ... View more

Writing ugly spaghetti code like this is a recipe for problems. DATA Count; SET Medications; BY Patient; IF First.Patient THEN DO; IF Medication = 'Aspirin' THEN DO; Count = 0; Count + 1; IF LAST.SubjectNumber; END; END; RUN; Now you can clearly see that the increment is part of the FIRST. block and therefore executed only once per patient. You also used a variable not included in the BY for your LAST. data count; set medications; by patient; if first.patient then count = 0; if medication = 'Aspirin' then count + 1; if last.patient; run;   ... View more