Yes, that makes sense.

Yes, with recreating the index I will use the following order

Non-clustered -> primary key -> clustered

So rebuilding of whole table will be performed last.

Thought?

4 non-clustered, 1 clustered and 1 primary key.

For bulk insert.

Thanks for the reply, I found that and managed to locate the files affected. But what is the error behind it? Not seems to be able to find the solution.

[Code]

select top 100 u58.deal_type_cde , ucb.*

from u58_deal_hist u58

join ucb_exp_hist ucb on u58.deal_id = ucb.deal_id

where ucb.batch_id = 297326

and u58.batch_id = 297325

and ucb.deal_id like 'IMGATD%'

[/code]

There are indexes...

This is what I get when trying to run the query

Msg 240, Level 16, State 1, Line 6

Types don't match between the anchor and the recursive part in column "String"...

Hi All

The query return in a table however I would like to return in 1 single row. eg "Project Lead;Developers"

Thanks

Yes, I am trying to stimulate a dead lock.

Thanks I think I have enough info.

Derek

Thanks for reply everyone. The posts are really help.

Thanks for the solution. It works perfectly. What is the function over (Partition BY) does? What is the different between Over(Partition By) and Over(Order by).

As a high level overview, what you'll be wanting to do s look at the exec plan, find the high percentage cost operations and see what you can do to...

Define performance? If you're concerned with the time that the query takes, use statistics time. If you're concerned wiht the amount of IO, use statistics read. If you're concerned with...

Thanks for prompt reply.