Are the posted questions getting worse?

L' Eomot Inversé (2/21/2013)

Right now, I'm feeling embarrassingly stupid. Blush

I was trying to put together a submission for QotD. But I thought I'd seen too many comments recently where people had run code where really if running that code was necessary to get the answer they need a very basic remedial SQL course, so I wanted to forulate the code in such a way that they couldn't reasonably run it because it would run for too long (or at least restrict the ability to run it to people who had enough personal resources to run it without getting fed up with waiting for a result, rather than running it on company machines because it would take a nasty chunk of machine power). At the same time, I wanted to be able to run it to make sure that my answer was correct. So I put together the first part of the code, which would generate data for a table or CTE (hadn't decided yet) that would take a silly amount of both CPU and disc activity to generate, in a form that I thought I could sensibly run on my rather underpowered machine and leave to complete while I went off to make and drink some coffee, make and eat some sandwiches, and pour and swallow a couple of jars.
I came back to my machine and found that the code had finished in less than two minutes - and failed horribly: the error message was

Msg 1101, Level 17, State 10, Line 12
Could not allocate a new page for database 'TEMPDB' because of insufficient disk space in filegroup 'DEFAULT'. Create the necessary space by dropping objects in the filegroup, adding additional files to the filegroup, or setting autogrowth on for existing files in the filegroup.

The problem was incredibly stupid ( Blush

). An intermediate result had to be either explicitly materialised as a (temp) table or spooled by the optimiser (the latter was the case in the version of the set-up that I tried). I had only a few 10s of gigabytes to spare. The intermediate result, if the machine had managed to generate and store it, would have had - as I discovered by doing some straightforwards arithmetic, which I ought to have done before setting it running Blush

- a little more than 2**48 (256 trillion - counting 1024 instead of 1000 as thousand) rows (actually 281475010265089, which is 1+2**25+2**48). w00t

Maybe that possible QotD will never see the day - I'll have to see if I can make it smaller without turning it into trivia (the crux was going to be a MERGE statement with three merge clauses, all of them with clause search conditions - maybe I should try to curb my liking for bizarre questions).