Actually, I only submitted the winning entry; Denali CTP3 SQL Server 2012 is the real winner. After having a blog posted here recently about running totals performance in Denali CTP3 SQL Server 2012, I happened to run across this T-SQL Challenge for doing running totals. On a lark, I posted a Denali CTP3 SQL Server 2012 solution. I fully expected it to be discounted (on a not released version of SQL, other solutions would be faster, heck, anything)… but this simple solution turned out to completely trounce all of the other solutions and it was selected as the top winner! The SQL Server 2012 solution came in with 529 reads, no writes… some of the other solutions that produced the correct answer (and thus became a “winning” solution) came in with MILLIONS of reads, and some with THOUSANDS of writes. Duration? The closest competitor came in at twice the time. Honestly, I’d hate to see what the performance is of some of the other solutions that produced the correct results, but were not performing well enough to make it into the top echelon.
So, for all those fine folks at Microsoft that implemented the ANSI windowing: thank you for what you did. This just goes to show that this can now enable you to write simpler, faster code quicker than ever before. Thanks!!!
Congratulations Wayne. You are a very good SQL DBA\Developer so I’m not surprised.
Reply
Thanks Jack… but it is really SQL Server 2012 that is the real winner here. The expanded capabilities within it is what allowed this code to win.
Reply