> I don't know. I'd say it's more "correct" in some abstract theoretical 
> sense, since the only reason you're grouping is to get unique rows. And it 
> works in the case you're trying. :-)
>
> If you wanted to know how many different types of events each experiment 
> yielded, you'd do something like
>   select expid, count(eventid) from exp group by expid
> or something like that. (Not gonna fire up SQL to try it.)
>
> But you see what I'm saying there.

Yeh I see what you're saying, I'll have to work on importing more realistic 
data as my test data is a bit artificial.  But the speed so far seems fine 
and I'm pretty confident that SQL CE will be capable of doing everything I 
need.  If I had known it was this easy to set up and work with .net I might 
have used it for some other projects too rather than struggling with 
filtering and sorting arrays and other data containers!

Post a reply to this message