Generally it is not necessary to worry about case when typing code and referring to variable names, dataset and library names etc., however a consistent coding style is always to be recommended.
This has two benefits: firstly it aids in readability, and secondly, there may be circumstances where the case may be an issue.
Consider the code:
data vars ; attrib a B c length = 4 ; a = 100 ; b = 200 ; c = 300 ; run ;
which runs and produces a dataset.
a | B | c |
100 | 200 | 300 |
Look at the properties of this dataset and you will see that the variable names are: a, B and c which is the case in which they were first referenced, and the system retains these.
Any subsequent code which references the second variable can use either b or B interchangeably.
If however the data is transposed:
proc transpose data = vars out = vars_t (rename = (_name_ = cvar)) ; run ;
the output dataset now has the original variable names stored as values in the cvar variable: these values are now case-sensitive, and inconsistent.
cvar | COL1 |
a | 100 |
B | 200 |
c | 300 |