Congratulations Leszek - that was a very close race in the end!

@Ramesh Saive Thanks for posting those stats - one thing though... the 100,000 unique values data is Dishdy's, not mine (minor point) . edit: removed incorrect statement

@Stefan_G, @dishdy I agree - the challenge is possible as it stands - we gain nothing from changing the rules now...

So,my final stats for 100,000 unique names (under parallelism) are horrible - let's hope the actual perf test data isn't that bad! Table 'TC50'. Scan count 7, logical reads 702, physical reads 0, read-ahead reads 0, lob logical reads 0

"

@Stefan_G I do : [http://www.sqlservercentral.com/Forums/Topic1059167-145-1.aspx][1] It descends into a bit of a scrap quite quickly - not really a technical debate so much as a fight. (Can't create a link for some reason) [1]: http://www.s

Agreed - good challenge - but boy it would stir up a storm with some people - a recent "discussion" on SSC about an interview question like this comes to mind... Anyway, my stats: Sample Data: Table 'TC50'. Scan count 13, logical rea