Doing transactional replication, set up with nosync initialization, it
seems to work great, but does not generate and does not use custom SPs
for each input/update/delete.

In the profiler I can see that, after I run a simple "update"
statement on the publisher, the same SQL gets run on the subscriber.

Which parameter of what call is it that determines whether I am using
this "direct update" business, and what are the implications, and
where (if anywhere!) is this discussed in BOL (or elsewhere), cuz I
can't find it!

Thanks.

Joshua Stern

On Mon, 22 Nov 2004 14:27:08 -0800, jxstern <jxstern (AT) nowhere (DOT) com>
wrote:
Found it (in my own scripts, of course!)

exec sp_addarticle @publication = N'myPUB', @article = N'tbl_foo',
@source_owner = N'dbo', @source_object = N'tbl_foo',
@destination_table = N'tbl_foo',
@type = N'logbased', @creation_script = null, @description = null,
@pre_creation_cmd = N'drop',
@schema_option = 0x00000000000000F3, @status = 0,
@vertical_partition = N'false',
@ins_cmd = N'SQL', @del_cmd = N'SQL', @upd_cmd = N'SQL',
@filter = null, @sync_object = null, @auto_identity_range = N'false'

Those "@ins_cmd" and such say "SQL", which echoes the SQL, rather than
use CALL, MCALL, or XCALL. See also BOL section, "Using Custom Stored
Procedures in Articles"

J.