Re: Popcon time-series data points
Hello,
On Sat, 2018-02-24 at 15:15 +0100, Lucas Nussbaum wrote:
> Hi,
>
> On 23/02/18 at 15:04 +0000, Nikolaos Alexopoulos wrote:
> > On Wed, 2018-02-14 at 13:57 +0100, Lucas Nussbaum wrote:
> > > On 14/02/18 at 13:43 +0100, Bill Allombert wrote:
> > > > On Wed, Feb 14, 2018 at 12:47:17PM +0100, Lucas Nussbaum wrote:
> > > > > (Adding popcon developers to Cc, and reordering paragraphs so
> > > > > that it's
> > > > > easier to follow)
> > > >
> > > > Hello Lucas,
> > > > Please note that popcon-developers@alioth is now debian-popcon@
> > > > list
> > > > s.
> > >
> > > ah, sorry, I missed that list.
> > >
> > > > > OK, I looked into how the popcon graphs are implemented.
> > > > >
> > > > > If nobody objects by 2018-02-21, I'll send you the output of
> > > > >
> > > > > select p.day, package, p.vote, p.old, p.recent, p.no_files,
> > > > > submissions
> > > > > from popcon_package pp
> > > > > join popcon p on (pp.id = p.package_id)
> > > > > join popcon_day pd on (pd.day = p.day)
> > > > > where in_debian order by p.day;
> > > >
> > > > The popcon graphs are only available for official packages.
> > > > Please make sure you only send the data for the packages for
> > > > which the graphs are available.
> > >
> > > Right, I think that's what 'in_debian' does in the above query.
> > > At
> > > least
> > > that's what 'select * from popcon_package where not in_debian;'
> > > makes
> > > me
> > > think.
> > >
> > > Lucas
> >
> > Kind reminder.
>
> I've just sent you a private mail with a link to the popcon data
> generated by:
> \copy ( select p.day, package, p.vote, p.old, p.recent, p.no_files,
> submissions from popcon_package pp join popcon p on (pp.id =
> p.package_id) join popcon_day pd on (pd.day = p.day) where in_debian
> and
> date_part('day',p.day)=1 order by p.day) To '/tmp/popcon.csv' With
> CSV
>
> (so you get one record per package per month)
>
> Could you please send us a pointer to your research when you are
> done?
>
> Thanks!
>
> Lucas
Thanks for the data! I will provide you with a link to our research as soon as we have some results.
Cheers,
Nikos
Reply to: