Andrew Sullivan wrote:
> On Fri, Mar 10, 2006 at 02:55:24PM -0600, David Durham wrote:
> 
>>Still open is whether or not Slony would do anything with the BLOB data 
>>even the insert event might have been captured.
> 
> 
> No.  That's what the problem is: it can't, because the lo_ functions
> use OIDs as keys, and no triggers would fire.

Actually, just to be sure, can a trigger on a (non-oid) column cause 
replication of the large object (whose oid is in the row) and then 
replication of the entire row?

Sorry to keep on this issue.  This is the last time I'll rephrase my 
question.  Thanks.


Dave