Guessing without metrics.

Try updating stats on all tables involved in the process with full scan.

cygne17_2 77208 (2/18/2014)

---

Which scenario do you recommend?

1- Backup the database, backup the transaction log to a new disk, shrink the transaction log file, schedule transaction log backup each hour.

2- Backup...

Chapter 4 - http://www.red-gate.com/community/books/accidental-dba. All about memory, memory contention.

Just bear in mind that performs the operation and then undoes it. If a lot of...

Objects aren't defined on files, they're defined on filegroups. If you have two files in a filegroup, all objects on that filegroup will be spread across both files.

If you want...

TomThomson (2/18/2014)

---

What happens if that nullable column containts the string '12345' or '42 ' or anything else which will test equal to the result of casting some integer to...

TomThomson (2/18/2014)

---

not clustered isn't a sensible option - a view has to have a clustered unique index before...

Other than what I just said one message back, not really. Too open ended, not enough information. Test under expected volumes and see.

Yup, that's it. Should be less overhead than an indexed view.

This site is for Microsoft SQL Server, not any generic SQL database. Maybe try http://www.dbforums.com, they should have a sybase sections

Why a view, rather than just an indexed computed column? It'll work, but....

Do you not have a column that is already unique in the table, eg the primary key columns?

So then you're going to need to dig into the backup app's logs and find the detailed messages. The message in the SQL error log is just a 'something happened'...

I already did.

GilaMonster (2/17/2014)

---

For 2005, define a computed column on ISNULL(<column that you want to be unique>,<the primary key column of the table>) and put a unique index on that...

It's a corrupt database. Ask your DBA to run a database consistency check.

Why clustered? This isn't typically where you would put the clustered index.