July 4, 2011 at 11:44 am
1) It is not a rumor - Denali will have more complete support for windowing functions, including those that allow for single-pass running totals.
2) Given that a flaw was found what, a two years or more, after the "quirky update" was first put out there how many incorrect results were calculated? And was is the next unknown flaw that has yet to be found? This isn't a bug in the engine that resulted in wrong answers but is now fixed so we are OK to proceed. You are rigging the system using incomplete knowledge and using mechanisms that are not (as has been proven time and again) guaranteed to provide the correct output. What is the next "rule" that will need to be added? How many people out there are using prior "rules collections" that are now invalidated but whom will not find out about this new flaw?
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
July 4, 2011 at 12:58 pm
TheSQLGuru (7/4/2011)
Given that a flaw was found what, a two years or more, after the "quirky update" was first put out there how many incorrect results were calculated?
So, what's the flaw, Kevin?
--Jeff Moden
Change is inevitable... Change for the better is not.
July 5, 2011 at 6:53 am
Lynn Pettis (7/3/2011)
Is there a link to the updated code with saftey checks? I haven't been on ssc as much as I'd like to be so I think I may have missed part of this exchange of knowledge.
See http://www.sqlservercentral.com/Forums/FindPost980118.aspx and http://www.sqlservercentral.com/Forums/FindPost981258.aspx for the logic behind the safety check process.
Wayne
Microsoft Certified Master: SQL Server 2008
Author - SQL Server T-SQL Recipes
July 5, 2011 at 9:27 am
Jeff Moden (7/4/2011)
TheSQLGuru (7/4/2011)
Given that a flaw was found what, a two years or more, after the "quirky update" was first put out there how many incorrect results were calculated?So, what's the flaw, Kevin?
The compute scalar one that Tom and Paul came up with a 'workaround' to address.
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
July 5, 2011 at 5:11 pm
Nope. The thing that Paul and Tom came up with is nothing more than an error checking method. It also happens to help guarantee functionality, but there was no flaw there and Paul didn't originally write it as a guarantee helper. Just a method for quickly doing the error checking instead of using a separate sproc to do the error checking.
--Jeff Moden
Change is inevitable... Change for the better is not.
Viewing 5 posts - 16 through 19 (of 19 total)
You must be logged in to reply to this topic. Login to reply