Re: Row pattern recognition

On 6/28/23 14:17, Tatsuo Ishii wrote:
> Small question.
>
>> This query (with all the defaults made explicit):
>>
>> SELECT s.company, s.tdate, s.price,
>> FIRST_VALUE(s.tdate) OVER w,
>> LAST_VALUE(s.tdate) OVER w,
>> lowest OVER w
>> FROM stock AS s
>> WINDOW w AS (
>> PARTITION BY s.company
>> ORDER BY s.tdate
>> ROWS BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING
>> EXCLUDE NO OTHERS
>> MEASURES
>> LAST(DOWN) AS lowest
>> AFTER MATCH SKIP PAST LAST ROW
>> INITIAL PATTERN (START DOWN+ UP+)
>> DEFINE
>> START AS TRUE,
>> UP AS price > PREV(price),
>> DOWN AS price < PREV(price)
>> );
>
>> LAST(DOWN) AS lowest
>
> should be "LAST(DOWN.price) AS lowest"?

Yes, it should be. And the tdate='07-08-2023' row in the first
resultset should have '07-08-2023' and '07-10-2023' as its 4th and 5th
columns.

Since my brain is doing the processing instead of postgres, I made some
human errors. :-)
--
Vik Fearing