Re: graphing time series data

2010/4/14 Louis-David Mitterrand <vindex+lists-pgsql-sql(at)apartia(dot)org>:
> On Wed, Apr 14, 2010 at 08:46:13AM -0700, Richard Broersma wrote:
>> On Wed, Apr 14, 2010 at 7:54 AM, Louis-David Mitterrand
>> <vindex+lists-pgsql-sql(at)apartia(dot)org> wrote:
>>
>> > Now, I'd like to make a graph of average prices per week, per
>> > id_product. As some prices don't vary much, distribution would not be
>> > ideal if I simply 'group by extract(week from p.modified)'.
>>
>> I created a view for a similar problem that I had.  Only I was
>> calculating the counts per day.  this query could be crafted to work
>> for you.
>>
>> CREATE OR REPLACE VIEW opendiscrepencydailycounts AS
>>  WITH opendays(day) AS (
>>          SELECT gs.day::date AS day
>>            FROM generate_series((( SELECT
>> min(discrepencylist.discstartdt) AS min
>>                    FROM discrepencylist))::timestamp without time
>> zone, 'now'::text::date::timestamp without time zone, '1
>> day'::interval) gs(day)
>>         )
>>  SELECT opendays.day, ds.resolvingparty, count(opendays.day) AS
>> opendiscrepancies
>>    FROM discrepencylist ds, opendays
>>   WHERE opendays.day >= ds.discstartdt AND opendays.day <=
>> LEAST('now'::text::date, ds.resolutiondate)
>>   GROUP BY opendays.day, ds.resolvingparty
>>   ORDER BY opendays.day, ds.resolvingparty;
>
> You confirm my modus operandi. I tried the following which seems to give
> me optimal price distribution:
>
> select w.week,count( p.id_price) from
>                (select generate_series(min(p.created_on),max(p.modified_on),'1
>                week') as week from price p) as w join price p on (p.created_on <
>                w.week + '7 days' and p.modified_on > w.week + '7 days') group by
>                w.week order by w.week
>
>          week          | count
> ------------------------+--------
>  2010-02-10 15:32:18+01 | 125369
>  2010-02-17 15:32:18+01 | 126882
>  2010-02-24 15:32:18+01 | 128307
>  2010-03-03 15:32:18+01 | 126742
>  2010-03-10 15:32:18+01 | 133596
>  2010-03-17 15:32:18+01 | 149019
>  2010-03-24 15:32:18+01 | 149908
>  2010-03-31 15:32:18+02 | 147617
>
> The rest should be easy from there!

I don't understand well. Why you don't use a function date_trunc(),

select date_trunc('week', created), count(*)
from price
group by date_trunc('week', created)

Regards
Pavel Stehuke

>
> Thanks for your input,
>
> --
> Sent via pgsql-sql mailing list (pgsql-sql(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-sql
>