it is no different under unix then linux -- what I think you are seeing is different behavior in the C runtime library.

I'd have to see your code. You need to use NULL indicators (only way to deal with NULL's in C) and do something like:


exec sql whenever sqlerror continue;
exec sql drop table t;

EXEC SQL WHENEVER SQLERROR DO die();

exec sql create table t ( x varchar2(10), y varchar2(20) );
exec sql insert into t values ( 'Not null', 'hello world' );
exec sql insert into t values ( 'Is null', null );
exec sql commit;

exec sql declare C cursor for select x, y from t;
exec sql open c;
for( ;; )
{
exec sql whenever notfound do break;
exec sql fetch c into :x:x_ind, :y:y_ind;
printf( "%s %s\n", (x_ind?"":x), (y_ind?"":y) );
}
exec sql whenever notfound continue;
exec sql close c;
exit(0);
}


See how in the printf we look at the indicator variables to see if the string is "NULL" or not null.

If you have not been doing this logic in the past, you have been what is known as "extremely lucky" not to have seg-faulted. Sounds like you are printing out a NULL pointer and your printf runtime is printing the word (null) -- I've seen that behavior with various C runtimes.


that program, when run, will print:

[tkyte@localhost t]$ ./t
connecting
connected
Not null hello world
Is null
[tkyte@localhost t]$


So -- use NULL indicators and check them -- there is no other way, I've a feeling you have been extraordinarily lucky -- but you might find you legacy code's behavior changes with a libc patch or anything like that!