it certainly seems like you would need to aggregate keys somewhere (if it's not the typo i assumed). that might just be possible in postgres these days (update: a quick google suggests it's in development for 9.3), but i am pretty sure not in mysql. on the other hand, for all i know, at this scale, they may be treating mysql as some kind of distributed hashtable and doing a lot more work in higher layers...
[update: I completely screwed up the maths; looks like this is not rows per tweet, but support data.]
> for all i know, at this scale, they may be treating
> mysql as some kind of distributed hashtable and doing
> a lot more work in higher layers...
If MySQL is just used as hash table, and all of the interesting stuff happens at higher layers, then talking about the number of rows isn't that meaningful.
[update: I completely screwed up the maths; looks like this is not rows per tweet, but support data.]