Re: Behavior of subselects in target lists and order by

Поиск
Список
Период
Сортировка
От John R Pierce
Тема Re: Behavior of subselects in target lists and order by
Дата
Msg-id 4F48242B.9010802@hogranch.com
обсуждение исходный текст
Ответ на Behavior of subselects in target lists and order by  (amit sehas <cun23@yahoo.com>)
Список pgsql-general
On 02/24/12 12:45 PM, amit sehas wrote:
> If we have a query of the form:
>
> Select *, (Select * FROM T2 WHERE p2 = T1.p1) FROM T1 ORDER BY 1 WHERE p3 = 75

ORDER BY has to be AFTER the WHERE clause.


is that query equivalent to...

Select t1.*, t2.* FROM T1 LEFT JOIN T2 on T1.p1=t2.p2 WHERE t1.p3 = 75 ORDER 1

?


>
> In the above query there is a subselect in the target list and the ORDERBY has an ordinal number which indicates
orderby column 1. Does this mean that the above query will return all results from T1 that match p3 =75 and all results
fromT2 that match p2 = T1.p1 for every match on T1 and order them all by the first column of T1 and T2 ? 

I'm not sure your query as written would even work.  if that inner
select was to return multiple rows, ugh.


> basically i am trying to determine if the order by clause has effect only on the tuples of the outer select or both
theouter and inner select. Or the results returned by the inner select are treated as if they are part of a single
tuplewhich includes the tuple from table T1 ? 

you can't treat a recordset as a single tuple regardless.   the ORDER BY
is always done on the data selected by the WHERE


> Is this an implementation specific behaviour or it conforms to the SQL standard ...?
>


I suspect its your proposed query that doesn't conform.

--
john r pierce                            N 37, W 122
santa cruz ca                         mid-left coast


В списке pgsql-general по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: Re: [PERFORM] Disable-spinlocks while compiling postgres 9.1 for ARM Cortex A8
Следующее
От: John R Pierce
Дата:
Сообщение: Re: Pull the Trigger and Where's the SQL?