Yes, you are right. The only thing I can think of is that the driver is not updating the data length from the SQLBindCol to let DBD::ODBC the data length has changed (from the trace file).
If the data length is SQL_NULL_DATA, then DBD::ODBC makes sure the field is set to undef, otherwise, it used the data retrieved (pushed) to DBD::ODBC, since I'm using SQLBindCol. That may be different than what isql does...since Nick already commented that it's using SQLGetData...DBD::ODBC only uses SQLGetData in rare circumstances. I'd check on your versions of FreeTDS and unixODBC... Regards, Jeff > -----Original Message----- > From: Dave Thorn [mailto:davet@;fysh.org] > Sent: Tuesday, October 22, 2002 12:16 PM > To: Jeff Urlwin > Cc: [EMAIL PROTECTED] > Subject: Re: :ODBC bug? > > > On Tue, Oct 22, 2002 at 12:11:43PM -0400, Jeff Urlwin wrote: > > Right, but I only see one row in the message, per query. > Shouldn't there be > > two? > > I think you're mis-parsing it. The DURATION and PERFORMER bits are in > different rows to each other. > > > | DURATION | > > | > > | PERFORMER | RICHARD G. MITCHELL > > two rows. I took the liberty of selecting less fields than with the > previous queries to try and make it slightly more readable, sorry, > should have mentioned that, hoped it'd show in the output. > > > | PERFORMER | RICHARD G. MITCHELL > > | > > | DURATION | > > two rows (other query) > > -- > dave thorn | [EMAIL PROTECTED] >
