Dear Folks
need help urgently on an update statement
This is the setup:

TABLE_A has lets say 100 rows
lets say TABLE_A it has 3 columns, CODE, DESCRIPTION and SECONDFIELD,
(all varchars).

TABLE_B has two columns, CODE, DESCRIPTION (both varchars)

I want to update TABLE_A.DESCRIPTION, with the value TABLE_B.DESCRIPTION
joined by field 'CODE' which is common in both.
However, I want to affect only the records of target TABLE_A, where
SECONDFIELD, has a scalar condition lets say '1' In other words
something like that

update TABLE_A a set a.description= (select b.DESCRIPTION from
TABLE_B b where a.SECONDFIELD='1' and a.CODE=b.CODE)

Now, target TABLE_A has lets say 50 records where SECONDFIELD='1' and
50 more where SECONDFIELD='2'

the statement is ok as far as syntax and it executes. If i execute
this, it does the following:

It says that ALL 100 records were affected.(Why?)

The 50 records where SECONDFIELD='1' get updated alright, the other 50
where
SECONDFIELD='2' become null.

If I execute the opposite

update TABLE_A a set a.description= (select b.DESCRIPTION from
TABLE_B b where a.SECONDFIELD='2' and a.CODE=b.CODE)

you guessed it, the 50 records where SECONDFIELD='2' get updated
alright, the other 50 where
SECONDFIELD='1' become null and lose their original values.
This statement should have updated (or affected) only the records
where the join is succesfull, in this case only 50 of them in either
case based on the 'where a.SECONDFIELD=...' clause, and not all,
(At least this is what happens in Oracle)
why does this happen?

I know I miss something, any quick ideas?
Thanks in advance