I can't see the attachment (security issues).

But in general you want to avoid using functions on columns, because it keeps SQL from using indexes (and likely stats) on that column.

For...

Sorry, I mis-worded my response.

I didn't see the need for the DealType in the Deals table, since it had its own table. Indeed, since its many-to-one, I don't how...

Table 1: Deals (PK DealId)

DealId int

DealTypeId int

Table 2: DealEvents (PK DealId)

DealId int

DealFamilyId (computed persisted '1')

Table 3: DealTypes (PK DealTypeId)

DealTypeId int

DealFamilyId int

I want to create a relationship between Table1 and...

Interesting. Thanks for the update.

A little confusing though, not sure what's going on there then. Thought originally it might have just been a mismatched name because of the...

On my SQL 2005 version and my SQL 2008 versions, it accepts sp_rename for that index name:

USE tempdb

CREATE TABLE dbo.nasty_index_name_test ( c1 int )

GO

CREATE NONCLUSTERED INDEX [<Name of Missing Index,...

What do you mean "doesn't work"? It won't even initiate the rename process?

If the tables have identity columns, you can use $IDENTITY as the column name.

If not, you might have to dynamically generate the code based on unique indexes on the table...

You can rename an index through SSMS; just backspace over the old name, and type in the new one.

No, not CPU (except perhaps a tiny amount periodically to check it for activity).

Connections themselves have some RAM overhead.

But since creating/destroying connections is...

User dbs will definitely work ok to a later SP.

Master will definitely NOT work; master MUST be exactly the same version.

Not sure about msdb, but apparently same rules apply as...

Please try this 🙂 :

SET DATEFORMAT dmy

SELECT CAST('20120524' AS datetime)

SELECT CAST('2012-05-24' AS datetime)

The universally accepted date format for SQL Server is "YYYYMMDD" (as in 20120524), not YYYY-MM-DD, which different installations will treat differently (and some will error out).

SQL itself won't allow you to restore an older master db (not sure about the model db, but it likely doesn't contain critical data anyway). Something else must be...

Partitioning could indirectly help performance, as follows:

-- make it possible/faster to rebuild/reorg current yr's data only;

-- each partition can be separately specified as uncompressed/compressed (SQL 2008+ only).

Exactly Paul; well put.