Thank you to each of you who responded with help.
Unfortunately, while testing a Python script, I found
the drawback to the geometric mean calculation:
If your sample size is in the hundreds of thousands,
it's easy to hit the hardware (processor) limitation
on the number size while multiplying t
This is probably not the most elegent solution, but here is something
that works in 3 SQL statements using user variables.
mysql> select @gmean := 1, @ct := count(*) from g_mean;
mysql> select @gmean := @gmean * test_val from g_mean;
mysql> select power(@gmean, (1/@ct)) as Geo_Mean;
See details
Interesting idea, Andrew. I don't know how to do what you're asking for,
but I'd be interested to see it added too. I'd never heard of geometric
means, but I can already think of several areas where they might be
useful. Thanks for sharing the information on how they're calculated!
Matthew Walker
