Are you sure it isn't easier to re-read the data?
If the file started as any form of text I might ask why you have characters you don't want.
You have several issues.
First, nowhere do you define NVARS and apparently it is not a variable in your data set.
Vname returns the name of the variable if you have a an array reference like Vname(array(i)) . If you use Vname(somevariable) it will return the literal text "somevariable" if that is the name of a variable in the set.
So your Vname does not have any reference to anything related to variables to return those names.
The Rename statement in a data step will not allow use of expressions. So Rename varname=newname would actually be renaming the variable Varname that you just created to newname.
You can use information to build a control data set to create rename statements, preferably for Proc Datasets that is intended to do things like change properties of datasets including variable names, formats, labels and informats.
A small working example:
/* create something to work with*/
data work.have;
input a b_2017 c_2017 $;
datalines;
1 2 A
4 5 B
;
run;
proc sql;
create table rename as
select name, tranwrd(name,'_2017','') as newname
from dictionary.columns
where libname='WORK' and memname='HAVE'
and index(name,'_2017')>0
;
quit;
data _null_;
set rename end=last;
if _n_=1 then call execute('proc datasets library=work nodetails nolist; modify have; rename');
call execute (name||'='||newname);
if last then call execute('; quit;');
run;
What this does: The Proc sql uses the metadata maintained by SAS to search in the data set indicated in the Libname and Memname (data set) that have variable names that have the text _2017 in the name. The Library and Memname values are stored in the metadata in upper case so you want to use that in the where clause. Caution: I am not searching for the "last 5 characters" just the presence of the string '_2017'. If you have that as valid text as part of the variable names in other positions I would recommend going back and rereading the data from start. The SQL also creates a new variable with the modified name removing _2017. Extreme caution: There is a chance, depending on YOUR data that removing that next from the name might result in the name of another variable that already exists in your data set. Have you checked for that? Example your data set as variable Measure and Measure_2017. If you attempt to rename Measure_2017 as Measure you will not get it as that variable already exists.
The data _null_ step uses the Call execute function to place statements in to the execution buffer. It starts with boilerplate to use proc datasets with your library , set some options to reduce output, and then select the data set to modify and then the instruction rename. the next call execute creates the rename instructions. The data set option END creates a temporary variable Last that is true when the observation the data step is processing is the last one. So we can provide the code to end the Proc data sets call.
For my example the log shows:
NOTE: CALL EXECUTE generated line.
1 + proc datasets library=work nodetails nolist; modify have;
NOTE: Writing HTML Body file: sashtml1.htm
1 + rename
2 + b_2017 =b
3 + c_2017 =c
4 + ;
NOTE: Renaming variable b_2017 to b.
NOTE: Renaming variable c_2017 to c.
4 + quit;
You should be able to compare those lines after the + with the Call execute statements.
Note that Proc Datasets is one of a few procedures that does "run group processing" and could have multiple Run statements in one procedure call. This procedure uses QUIT; to end the procedure.
Click image to register for webinar
