To test your code there are various ways. Unit test are the most common one, not really a SAS concept though, a programming methodology more so.
*My* method is to run a PROC FREQ on the input variable. I look for all of the different unique cases that are possible and create small dataset with those values only. You can pipe your PROC FREQ output to a data set to help with this. Then you test it on the small data set where you can verify all data points. Once it's verified then you apply it to your big dataset.
You can also run a PROC FREQ on the input variable and compare it to the output but if you have 4million unique values.
proc freq data=want;
table inputVariable * outputVariable / out = check;
run;
@GN0001 wrote:
Hi Reza & all, What I asked was when we apply a function to a data set, for example a scan function on a values of a variable, how can we make sure our function has modified all the values?
I usually eyeball it and a little of log can help. Is there any other approach?
please advise me.
respectfully,
blue blue
... View more