Under BT2002 I had a sql adapter which called a stored proc which would
return 200 rows at a time for processing by BT.   The sql adapter under
BT2002 would retrieve these 200 rows in one call and then spawn multiple
instances of my BT2002 orchestration to process a row at a time.  Under
BT2004, however, I had to change my stored proc to return only one row at a
time in order to use the fields as promoted fields in my orchestration.   No
w
my through put is reduced because the adapter only runs every second and I
can't have parallel processing like I did using BT2002 ( unless I create
multiple adapaters ).   Any suggestions on processing multiple rows in
parallel like under BT2002 ?

[ Post a follow-up to this message ]

Have you tried setting the flag on the SQL receive port so that it keeps
retrieving rows until there are none left?

Alternatively, you can retrieve the 200 rows as a batch and then have a
pipeline spliit them up.

--
McGeeky
http://mcgeeky.blogspot.com


"horochj" <horochj@discussions.microsoft.com> wrote in message
news:7080CAD7-A706-44B9-AD7C-3948519CAB73@microsoft.com...
> Under BT2002 I had a sql adapter which called a stored proc which would
> return 200 rows at a time for processing by BT.   The sql adapter under
> BT2002 would retrieve these 200 rows in one call and then spawn multiple
> instances of my BT2002 orchestration to process a row at a time.  Under
> BT2004, however, I had to change my stored proc to return only one row at
a
> time in order to use the fields as promoted fields in my orchestration.
Now
> my through put is reduced because the adapter only runs every second and I
> can't have parallel processing like I did using BT2002 ( unless I create
> multiple adapaters ).   Any suggestions on processing multiple rows in
> parallel like under BT2002 ?

[ Post a follow-up to this message ]