Re: Invalid integrity support

From: Martin Haltmayer <Martin.Haltmayer_at_0800-einwahl.de>
Date: 2000/05/10
Message-ID: <3919389A.B2B76BBA@0800-einwahl.de>#1/1

Confirmation for NT4:

SQL>
SQL> drop table tmp1 cascade constraints;

Table dropped.

SQL> drop table tmp2 cascade constraints;

Table dropped.

SQL>
SQL> create table tmp1 (

  2  	     a number
  3  	     , constraint tmp1_pk primary key (a)

  4 )
  5 /

Table created.

SQL>
SQL> create table tmp2 (

  2  	     b number
  3  	     , constraint tmp2_tmp1_f foreign key (b) references tmp1 (a)

  4 )
  5 /

Table created.

SQL>
SQL> insert into tmp1 values (235);

1 row created.

SQL>
SQL> insert into tmp2 values (235);

1 row created.

SQL>
SQL> create index tmp2_b on tmp2 (b) reverse;

Index created.

SQL>
SQL> delete from tmp1;

1 row deleted.

SQL>
SQL> select * from tmp1;

no rows selected

SQL> select * from tmp2;

         B                                                                      
----------                                                                      
       235                                                                      

SQL>
SQL> spool off

Martin

Evgeny Faddeenkov wrote:
>
> Hello all,
>
> recently, I was impressed by some integrity
> rules invalidation. You can test it (by a very simple
> method) on your platform by applying the
> following code:
>
> create table tmp1(a number primary key)
> create table tmp2(b references tmp1)
> <inser records in both tables>
> create index tmp2_b on tmp2(b) REVERSE
> <delete records from tmp1>
> select * from tmp2
>
> I tested this example on Sparc Solaris and
> Linux with 8.0.5.0/ 8.0.5.2.1/ 8.1.5.0 RDBMS.
> And I saw that I can delete records on the
> master table without any effect on slave.
>
> So you'll receive an incorrect data integrity!
>
> It work's so only with REVERSE index.
>
> Sincerely yours,
>
> Evgeny Faddeenkov
Received on Wed May 10 2000 - 00:00:00 CDT