Caught up with few Issues, Advice is highly appreciated:
We are trying to improve the performance of our query by adding some indexes amd eliminating some unwanted stuff. When I run the old and the new query in o together and check the execution plan, I could see Cost relative to batch, the old query
is higher then the new query,I think that is a good sign but what I want to know is how true it is. The reason for asking is I have read it is only estimate and not neccessarily be true.
And also , While testing I would like to clear the server cache, I am using dbcc dropcleanbuffers command for that, Does that clear the server cache or is there any other Command?
Is it good to have Clustered index on identity field, it also happens to be the PK of the table.Are there any disadvantages in making identity field as PK and have clustered index on it?
While doing Transactional Replication, if we happen to add a new table at publisher, does that not replicated on to the subscriber unless a snapshot is generated again, if not, is there anyway we can generate the snapshot only for that table and replicate
it,because I do not want to generate the snapshot for the entire DB.
View Complete Post