SAS: Select string with blank at the end - string

Question

I have a problem while selecting data from an Oracle table. Some values are R, RJ, and so on. However when I run the following query I just get RJ:
proc sql noprint;
SELECT col
FROM myoracletable
WHERE col IN('R','RJ')
;
quit;
So I checked the value R at Oracle:
select distinct rawtohex(col) as col
from myoracletable;
The result for R is 5220. So the string is R[blank]. I modified my SAS program like this:
proc sql noprint;
SELECT col
FROM myoracletable
WHERE col IN('R ','RJ','5220'x)
;
quit;
However the entries with R are still not selected.
How can I solve this issue without trim or compress the string?

Answer 1

SAS uses only fixed length strings and strips trailing blanks to make it work. So it looks like SAS is pushing 'R' instead of 'R ' into the database when it converts your query for you.
You need to write the Oracle query directly instead. So instead of using implicit syntax like:
libname myora oracle ... schema=myoraschema ... ;
proc sql ;
SELECT col
FROM myora.mytable
WHERE col IN('R ','RJ')
;
You should use explicit syntax like this:
libname myora oracle ..... ;
proc sql ;
connect using myora ;
select * from connection to myora
(SELECT col
FROM myoraschema.mytable
WHERE col IN('R ','RJ')
)
;
The long term solution is to fix the Oracle table to NOT store trailing blanks. You might need to redefine the variable as VARCHAR(2) instead of CHAR(2).

Answer 2

I wasn't able to reproduce this behaviour within SAS itself:
data have;
input mystr :$2.;
cards;
R
RJ
;
run;
proc sql;
select mystr from have where mystr in ('R', 'RJ');
quit;
This selects both values. So one option might be to run an initial query to copy part of the Oracle table into a temporary dataset in your SAS work library, then run another query on that.
You might also get different results if you use a pass-through query to access your Oracle table, but I can't test that.