On 12/2/2016 12:19 AM, Job wrote:
> We need to archive this data to generate report and analysis views.
on top of what Stephen said, I'd do as much preprocessing, and
preaggregation as possible to simplify and speed up those
report/analysis views.... for instance, storing min/max/mean rollups
per hour, per day, per month could mean the difference between a billion
row seq scan that takes hours and a few second query to graph a given
time range.
--
john r pierce, recycling bits in santa cruz