Miles Egan

Did we decide to leave this as is for the 3.0 release?

Thanks for your work on this Aaron. I got the same impression from the profiling I did earlier. There's not a single bottleneck in Arel, there's overhead spread...

I'm inclined to agree with Neeraj here. For the sake of simplicity and maintainability it's better to ask people to avoid used reserved SQL words as field names...

The thing that makes it difficult is that the order by parameter can be any sql expression, so you'd have to try to parse it, not just throw quotes around it.

Any further thoughts on this? Is the patch ok or should we take a different approach?

Daniel I think it's worthwhile to run each of your benchmark tests at least 10 times and take some averages. When I do this on my machine I get a fairly wide sp...

I did basically the same thing Daniel. It seemed to shave a little overhead off but not much.

As far as I can tell there's not a single big bottleneck in the Ar...

Ok I'll see if I can put together a good patch for that.

I was able to shave a little of the overhead off the arel code by explicitly defining all the methods that are called via method_missing in arel/lib/arel/engine...

I think what you're seeing here is a MySQL limitation:
http://dev.mysql.com/doc/refman/5.1/en/numeric-types.html

The largest value an int field can contain is 2...