On Tue, 31 Dec 2013 02:23:30 +,
Brent Wood wrote:
> This should help... In each temporary table convert the time parts to
> a timestamp, then create an index on each of these, then join on the
> timestamp.
[...]
Thank you, these were very useful suggestions, and lead to a very
efficient ta
This should help...
In each temporary table convert the time parts to a timestamp, then create an
index on each of these, then join on the timestamp.
ALTER table mmc add column timer timestamp without time zone;
UPDATE mmc set timer = (utc_year || '-' || utc_month || '-' || utc_day || ' '
||
On Mon, Dec 30, 2013 at 07:53:06PM -0600, Seb wrote:
> Given that the process involves a full join, I'm not sure I can do this
> in chunks (say breaking down the files into smaller pieces). Any
> suggestions would be greatly appreciated.
First, what I would probably do is merge the two files outsi
Quick thoughts:
On both tables:
Convert your date-time varchar fields into a single epoch/integer field.
Create an index of that epoch/integer field.
David J.
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/bulk-loading-table-via-join-of-2-large-staging-tables-tp57
