I agree that a builtin for std dev is a good idea. that said, you can achieve this easy in one pass, just use:
select sum( pow(col,2) ) as totsqr, sum( col ) as tot, count(1) as n, pow( (n*totsqr - pow(tot,2) )/(n*(n-1)), 0.5) as stddev
from .... Matt Pestritto wrote:
Hi.Are there plans to write a standard deviation aggregate function ? I had to build my own which translated into multiple hive queries. While it works, a build-in function would have been much easier.Thanks -Matt
