Martin Neimeier wrote:
> some additional informations:
> - if i execute the subselect alone, it works fine !.
> - The same select statement works with sybase and oracle, so i think its a legal statement.
> - After reading in the sql2-standard, i have found nothing which restricts unions in sub-selects.
Create a temp table (I did it this way):
select x from table where x in (select A union select B);
create temp table t1 (x type_of_x; );
insert into t1 select A union select B;
select x from table where exists (select t1.x from t1 where t1.x=table.x);
Using exists instead of in circumvents another restriction of PostgreSQL.
Tom Lane said, these bugs would be addressed during the query tree reorganization (7.2, in 2001)
> (I don't want to use another rdbms ... i want to use PostgreSQL :-)))))
> If somebody has a workaround, then i am the happiest person for the day.
Could be ... if you can live with this ...
In response to
pgsql-bugs by date
|Next:||From: Ludovic LANGE||Date: 2000-09-25 19:55:13|
|Subject: PostgreSQL crashes using distance operator whith records where 'point' data type set to null|
|Previous:||From: Tom Lane||Date: 2000-09-15 19:20:57|
|Subject: Re: libpq bug |