Re: Populating missing dates in postgresql data
David G. Johnston <david.g.johnston@gmail.com>
From: "David G. Johnston" <david.g.johnston@gmail.com>
To: Vincent Veyron <vv.lists@wanadoo.fr>
Cc: "Lavrenz, Steven M" <slavrenz@purdue.edu>,
"pgsql-general@postgresql.org" <pgsql-general@postgresql.org>
Date: 2015-03-27T15:33:36Z
Lists: pgsql-general
On Fri, Mar 27, 2015 at 3:41 AM, Vincent Veyron <vv.lists@wanadoo.fr> wrote: > On Thu, 26 Mar 2015 00:25:09 +0000 > "Lavrenz, Steven M" <slavrenz@purdue.edu> wrote: > > > I have a second table (TABLE B) with all of the object_ids and channels > that are supposed to be reporting in each day. For cases where a certain > channel does not check in, I want to add a column that indicates the comm > failure. > > Not sure if your context allows it, but if you can change your program's > logic, it might be easier to : > > -add a boolean field (e.g. 'checked') to table B, set to false > -whenever an object checks in do 'update table B set checked = true where > object_id = X and channel = Y' > -run a cron job once a day that > -runs 'select * from B where checked = false' and stores results > somewhere > -resets B with 'update B set checked = false' > > This is a performance optimization that I would avoid at nearly any cost, and there are likely better ways to limit the processing scope without having to trust the a cron job runs daily in order to not lose data. Now, there are other points of failure here that are of a similar nature already but still adding one for (pre-mature...) optimization doesn't seem like a good move. David J.