[Rd] BUG in RODBC() on Ubuntu 64amd : rows_at_time=1024

[Ken-JP]

R 2.8.1
Ubuntu 64amd
RODBC 1.2-5

---------------------------------

This was a hard-to-track bug for me, because sqlQuery() did not fail -
instead, it looked like it returned some good data and some NA where there
should have been numbers and 0 where there should have been NAs.

I did:

(a) sqlQuery( o, qry, as.is=TRUE, errors=TRUE, rows_at_time=1024 );
(b) sqlQuery( o, qry, as.is=TRUE, errors=TRUE );

My qry was a bit complex, with an outer join, so the first thing I did was
check from a win32 bit machine and check on 64amd with the mysql client
command-line.  Both worked fine.

Both (a) and (b) work fine from a XP Pro 32-bit R 2.8.1 installation.

On the amd64, when I tightened up my query to return less than 1024 lines,
(a) returned the correct result.  Over 1024, and (a) still returned the
right number of rows, but the results were incorrect.  Since I don't need
the rows_at_time argument, my workaround is to just use (b).

I wasn't sure where to post this, so I apologize in advance if this is the
wrong place.

- Ken



-- 
View this message in context: http://www.nabble.com/BUG-in-RODBC%28%29-on-Ubuntu-64amd-%3A-rows_at_time%3D1024-tp23019424p23019424.html
Sent from the R devel mailing list archive at Nabble.com.