Roz King just wrote an interesting article on binning data (a common data analytics step) in a database. He compares a case-based approach (where the bin divisions are stuffed into code) with a join based approach. He shares code and timings.
Best of all:
rquery gets some attention and turns out to be the dominant solution at all scales measured.
Here is an example timing (lower times better):
So please check the article out.