Hello,
Here is a test I did on 10gR2:
--local DB
create or replace view v1 as select count(*) c1 from t1@l1;
select * from v1;
--result: 0
--remote DB
insert into t1 values('a');
commit;
insert into t1 values('a');
commit;
--local DB
select * from v1;
--result: 1
select * from v1;
--result: 2
Results of the last two queries are inconsistent (the cause of it is explained in docs) - what bothers me is that if I change:
select * from v1;
to:
select count(*) from t1@l1;
then each query gives same (consistent results).
What's the difference (in the context of read consistency in distributed db) between issuing query using a view based on a remote query and issuing the same remote query directly?
Regards,
Andrzej