Re: UNION CASTING .

Go on 'hacking the datadictionary' you will loose support, and please stop asking questions here. If you think it is a 'serious bug' why don't you develop your own 'perfect RDBMS' where you can safely 'hack the datadictionary'.

Regards,

Sybrand Bakker, Oracle DBA

On Fri, 27 Apr 2001 16:25:28 +0100, "Maze Control - Terminal 23315 - Central 2" <trw-sinterface_at_pluto.gwy> wrote:

>I beg to differ.
>I consider this to be a serious bug in the database, however I do not really
>wish to get into a tiff over whether it is or not. All I am looking for is a
>method of hacking the data dictionary to correctly reflect the fact that
>nulls are unobtainable from this UNION .
>
>Thanks,
>Chris
>"Steve Long" <steven.long_at_erols.com> wrote in message
>news:9cbuff$q5b$1_at_bob.news.rcn.net...
>> there are simple views and complex views which accounts for the
 differences
>> you have observed.
>>
>> "Aparna" <bliss_is_ignorance_at_hotmail.com> wrote in message
>> news:3ae9755c.0_at_news.tm.net.my...
>> > oh one more though... we cannot perform DML's on views based on
 unions...
>> > the DML's can be performed only on the tables....
>> >
>> > regards,
>> > aparna
>> >
>> > "Steve Long" <steven.long_at_erols.com> wrote in message
>> > news:9cbof3$oln$1_at_bob.news.rcn.net...
>> > > my hunch is that the constraints are applied at the table level (as
 they
>> > > should be) rather than the view level. where did you create a not
 null
>> > > contraint on the "view column" ? if you constructed something similar
 as
 an
>> > > updateable view, a null value would result in a violation of the table
>> > > constraint. the dictionary is correct.
>> > >
>> > > "Maze Control - Terminal 23315 - Central 2" <trw-sinterface_at_pluto.gwy>
 wrote
>> > > in message news:0PbG6.12036$_W2.12062_at_news.indigo.ie...
>> > > > Can anyone help me with this ?
>> > > >
>> > > > Try the following SQL.
>> > > > --create table table1 (col1 integer not null)
>> > > > --create table table2 (col1 integer not null)
>> > > > --create view view1 as select * from table1
>> > > > ---union all select * from table2
>> > > >
>> > > >
>> > > >
>> > > > if you then do select column_name,nullable from user_tab_columns
 where
>> > > > table_name='VIEW1'
>> > > > the result is that col1 can contain nulls, which is incorrect.
>> > > >
>> > > > Try it on your own system - happens at least on Dynix and Win2000.
 V817
>> > > >
>> > > > Does anyone know how to hack the data dictionary to force this to be
>> > > > not-null ? (i.e. correct )
>> > > >
>> > > > Creating another view on top of this (AS SELECT COL1 from VIEW1
 WHERE
 COL1
>> > > > is NOT NULL )
>> > > > does not work .
>> > > >
>> > > > This is a small case of a much larger production issue. The view
 cannot
 be
>> > > > replaced with a table.
>> > > > TIA !
>> > > >
>> > > >
>> > > >
>> > >
>> > >
>> >
>> >
>>
>>
>
  Received on Fri Apr 27 2001 - 11:17:41 CDT