Connect Digest : 2009-11-25
Can we improve on CHECKSUM()?
Mike C# is not happy with the reliability of CHECKSUM() and asks for a better function that has less (or no) chance of collisions.
#513376 : A Better Collision-Free Hash Function for Comparing Rows of Data
Service termination and fatal exceptions are bad, m'kay?
Victor Lobanov found an easy way to crash SQL Server, by using obscure but legitimate values as terminators in a BULK INSERT statement.
#513252 : Certain values for ROWTERMINATOR field in BULK INSERT crash SQL Server 2008 SP1
And Jeremiah Peschka found another bug involving parallel queries, this time leading to a fatal exception (C000005 EXCEPTION_ACCESS_VIOLATION).
#514477 : Parallel plans combined with PIVOT and UNION ALL produce fatal exception
SSMS has bugs? Say it ain't so!
Jonathan Kehayias pointed out two bugs with the Database Mail Wizard, which seem to be tightly correlated, even though one has been closed as Won't Fix, and one is active but allegedly a duplicate of an existing work item.
#207602 : Max of 1 Database Mail Wizard open at a time
#463953 : SSMS Configure Database Mail does not setup mail correctly in some circumstances
And earlier this week, I was futzing around with the partition wizards that finally appeared in the 2008 version of SSMS. While it's great to see them, it is too bad that, like the Policy-Based Management wizards, they have similar restrictions on valid syntax that is accepted for things like date literals.
#513997 : SSMS : Create a Partition Wizard does not accept YYYYMMDD
I believe that Mike C# is our very own Michael Coles … http://sqlblog.com/blogs/michael_coles/default.aspx