topic Why are there two ways of renaming variables? And which method is the preferred method? in SAS Programming

Why are there two ways of renaming variables? And which method is the preferred method?

Nietzsche — Thu, 13 Oct 2022 05:03:58 GMT

In the textbook (Elliot - SAS essentials pg), it showed us two ways of renaming variable(s).

Method 1, with parenthesis

DATA NEW; SET TRANSPOSED (RENAME (INFO1=T1 INFO2=T2 INFO3=T3 INFO4=T4 INFO5=GENDER));
RUN;

Method 2, without parenthesis

DATA NEW; SET TRANSPOSED;
RENAME INFO1=T1 INFO2=T2 INFO3=T3 INFO4=T4 INFO5=GENDER;
RUN;

Is there any reason to used one method over the other one? or they are both equal in the eyes of SAS programming?

Re: Why are there two ways of renaming variables? And which method is the preferred method?

Kurt_Bremser — Thu, 13 Oct 2022 05:22:08 GMT

The dataset option allows renaming for each individual dataset (SET and MERGE allow multiple datasets) before the PDV is built, while the RENAME statement works in the context of the PDV.

Re: Why are there two ways of renaming variables? And which method is the preferred method?

Patrick — Thu, 13 Oct 2022 06:26:54 GMT

For your method 1 the correct syntax needs to include an equal sign.

DATA NEW; SET TRANSPOSED (RENAME=(INFO1=T1 INFO2=T2 INFO3=T3 INFO4=T4 INFO5=GENDER));
RUN;

Re: Why are there two ways of renaming variables? And which method is the preferred method?

Nietzsche — Thu, 13 Oct 2022 08:55:15 GMT

note to myself:

PDV = Program Data Vector

Re: Why are there two ways of renaming variables? And which method is the preferred method?

Kurt_Bremser — Thu, 13 Oct 2022 08:59:54 GMT

Oh yes, PDV is such a common acronym among SASians that we often overlook it may be confusing to beginners.

Re: Why are there two ways of renaming variables? And which method is the preferred method?

Tom — Thu, 13 Oct 2022 13:29:19 GMT

Those are not really equivalent.

The equivalent code using the RENAME= dataset option would be with the option on the output dataset.

The way you have it the names of the variables during the execution of the data step will be the NEW names.

You can use the RENAME= option in other code than data steps, such as PROC steps.

In general you should only use the RENAME= dataset option when you NEED to use it.

Here is an example where you need to use the RENAME= dataset option. There was a recent question where someone needed to match FID in one dataset to FAMID in another dataset. You could use the RENAME= dataset option on one of the input datasets to make that happen.

data want;
  merge dads(rename=(fid=famid)) kids;
  by famid;
run;