Update no where actually works

Question

Hi,
I just came across this update with no where clause and wondered why it actually works (And when it won't work)
update t1 set y = (select y from t2 where x = t1.x);Here's an example to run:
SQL> select *from v$version
BANNER                                                         
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
PL/SQL Release 10.2.0.4.0 - Production                         
CORE     10.2.0.4.0     Production                                     
TNS for IBM/AIX RISC System/6000: Version 10.2.0.4.0 - Productio
NLSRTL Version 10.2.0.4.0 - Production                         
5 rows selected.
SQL> create table t1 (x number not null, y number null)
Table created.
SQL> insert into t1 (x,y) values (1,null)
1 row created.
SQL> insert into t1 (x,y) values (2,null)
1 row created.
SQL> insert into t1 (x,y) values (3,null)
1 row created.
SQL> insert into t1 (x,y) values (4,null)
1 row created.
SQL> create table t2 as select x, 2*rownum y from t1 where x < 4
Table created.
SQL> select * from t1
         X          Y
         1          
         2          
         3          
         4          
4 rows selected.
SQL> select * from t2
         X          Y
         1          2
         2          4
         3          6
3 rows selected.
SQL> update t1 set y = (select y from t2 where x = t1.x)
4 rows updated.
SQL> commit
Commit complete.
SQL> select * from t1
         X          Y
         1          2
         2          4
         3          6
         4          
4 rows selected.Regards
Peter

Answer 1

I have modified your test case. Put 0 instead of null in T1.
SQL> create table t1 (x number not null, y number null);
Table created.
SQL> insert into t1 (x,y) values (1,0);
1 row created.
SQL> insert into t1 (x,y) values (2,0);
1 row created.
SQL> insert into t1 (x,y) values (3,0);
1 row created.
SQL> insert into t1 (x,y) values (4,0);
1 row created.
SQL> select * from t1;
         X          Y
         1          0
         2          0
         3          0
         4          0
SQL> create table t2 as select x, 2*rownum y from t1 where x < 4;
Table created.
SQL> select * from t2;
         X          Y
         1          2
         2          4
         3          6
SQL> update t1 set y = (select y from t2 where x = t1.x);
4 rows updated.
SQL> commit;
Commit complete.
SQL> select * from t1;
         X          Y
         1          2
         2          4
         3          6
         4Now you can easily notice that your update has modified all the rows in t1. And the in the row with X=4 the Y value is set as null as no match was found.
To avoid this situation you need to use a WHERE clause in your update to update only the matching records.
SQL> update t1 set y = 0
  2  /
4 rows updated.
SQL> select * from t1
  2  /
         X          Y
         1          0
         2          0
         3          0
         4          0
SQL> update t1 set y = (select y from t2 where x = t1.x)
  2   where exists (select y from t2 where x = t1.x)
  3  /
3 rows updated.
SQL> select * from t1
  2  /
         X          Y
         1          2
         2          4
         3          6
         4          0Now you can see for the row where X=4 the Y value is un touched.
Thanks,
Karthick.