Update of a table, but involving a join

Question

Hello,
I have a problem that I have not run into before, but it seems to be something commonplace and I thought I would post it here.
I have a source table and a target table, and want to copy out only the changes in the source table to the target table. I divided it into three parts, with a unique row being identified by the primary key columns.
- rows in the source not present in the target - these I simply insert into the target table
- rows in the target that have been deleted in the source - these I set a status flag to mean inactive
- rows that are common, but have had one or more non-key columns changed - these I copy out the non-key column values
I have run into problems with the last one. Perhaps it is my knowledge of Oracle syntax that inadequate, though one person told me that it could not be done.
I have tried some SQL variations, and here is something like what I tried. I have removed proprietary information.
update target
set target.column1 = common_rows_with_changes.column1, target.column2 = common_rows_with_changes.column2
from target,
select source.* from target, source
where target.primary_key_column = source.primary_key_column
and (target.column1 <> source.column1 or target.column2 <> source.column2 )
) common_rows_with_changes
where target.primary_key_column = common_rows_with_changes.primary_key_column
Thanks,
Regards,
Srini

Answer 1

update target
set ( target.column1 , target.column2 ) = (
select source.column1, source.column2
from source
where target.primary_key_column = source.primary_key_column
and (target.column1 <> source.column1 or target.column2 <> source.column2 )
where exists (
select null
from source
where target.primary_key_column = source.primary_key_column
and (target.column1 <> source.column1 or target.column2 <> source.column2 )
)