[Teammetrics-discuss] UDD Statistics

[Andreas Tille]

On Mon, Jul 11, 2011 at 11:52:46PM +0530, Sukhbir Singh wrote:
> I have been going over upload_history.py and create_names_preferred.sql.
> 
> Much of the work is handled by: active_uploader_per_year_of_pkggroup;
> seems like a lot of JOIN(s)!

That's NOT a lot of JOIN(s)! :-)

> You mentioned an alternate approach to
> this and since we are doing it from scratch, even if it is extra work,
> it makes sense to implement that... Thoughts?

IMHO this is a reasonable approach and I'd suggest to take it over
also for the liststats.  The main difference in the principle is that
crosstab (from postgresql-contrib) is used in contrast to the approach
in liststats where I created a query string with string operations.
While the listarchive table structure is way simpler than UDD and
does not even that few JOIN(s) :-)  I would recommend using crosstab
here as well.

If you might come up with some simpler SQL query I would not mind
but I doubt that you can make it much simpler.

Kind regards

      Andreas.

-- 
http://fam-tille.de