The SQL to review distribution across all segments
Post date: Oct 31, 2014 2:8:24 AM
select gp_segment_id, count(*) , abs(count(*) - avg(count(*)) over(order by gp_segment_id rows between unbounded preceding and unbounded following))/count(*) skew
from bloattest
group by gp_segment_id
order by gp_segment_id
limit 100;