The problem is solved. The coalesce() function was the solution. Thank you
very much for your input.
With best regards
Horst Düster
---------------------------------
Dr. Horst Düster
GIS-Koordinator /Amtschef Stv.
Kanton Solothurn
Amt für Geoinformation
Abteilung SO!GIS Koordination
Werkhofstr. 65
CH-4509 Solothurn
Tel.: ++41 (0)32 627 25 32
Fax: ++41 (0)32 627 22 14
horst.duester@bd.so.ch
www.sogis.so.ch
-----Ursprüngliche Nachricht-----
Von: Dawid Kuroczko [mailto:qnex42@gmail.com]
Gesendet am: Freitag, 8. Juli 2005 11:08
An: Düster Horst
Cc: pgsql-admin@postgresql.org
Betreff: Re: [ADMIN] where with NULL values are not selected
On 7/8/05, Düster Horst <Horst.Duester@bd.so.ch> wrote:
> I try to join two tables whereas some column values do have NULL values
with
> the following query:
>
> select table1.column from table1, table2 where table1.column=table2.column
>
> table1.column and table2.column may have NULL values. The problem is that
> these columns where not selected. Does there exists any solution to
> select/join the NULL value colums also.
A hint:
qnex=# SELECT 'ok' WHERE NULL=NULL;
?column?
(0 rows)
In other words -- NULL is not equal to NULL. NULL is not a value, NULL
is a state. If you want to join those columns, you cannot use NULL
as a joining key. It's the way SQL works.
Anyway, a quick-and-dirty solution might be:
select table1.column from table1, table2 where
coalesce(table1.column,-1)=coalesce(table2.column, -1);
...assuming column doesn't have '-1' value.
A better solution would be to rethink your design (don't use NULLs).
Incientally, instead of using this syntax:
select table1.column from table1, table2 where table1.column=table2.column;
...try using explicit inner joins:
select table1.column from table1 JOIN table2 USING(column);
Regards,
Dawid
