What is the Final Table equivalent for Merge?

Question

I'm trying to use merge into to insert only new records. I would like to collect the IDs for new records that got inserted and also the IDs for duplicate records that get ignored.

Here's the create statement for the table:

drop table SSZ_ME_MIS.test_update_table;
create table ssz_me_mis.test_update_table (
    ID_col int not null generated always as identity, -- Primary Key
    val_col_1 int not null,
    val_col_2 varchar(255) not null,
    constraint pk_test_update_table primary key (ID_col),
    constraint uq_test_update_table unique (val_col_1, val_col_2)
);

and then, to populate some initial values:

insert into ssz_me_mis.test_update_table (val_col_1, val_col_2)
select *
from (values 
    (231, 'Value 1'),
    (481, 'Value 2'),
    (813, 'Value 3')
);

So, finally, I'd like to try to do this sort of insert:

select ID_col from final table (
    merge into ssz_me_mis.test_update_table t using (
        select *
        from (values 
            (231, 'Value 1'),
            (481, 'Value 2'),
            (513, 'Value 4')
        )
    ) as s (val_col_1, val_col_2)
    on
        t.val_col_1 = s.val_col_1
        and t.val_col_2 = s.val_col_2
    when not matched then 
        insert (val_col_1, val_col_2)
        values (s.val_col_1, s.val_col_2)
    else
        ignore
);

Is there some way to accomplish this?

Answer 1

Something like this will run on Db2 LUW (assuming you are using ORGANIZE BY ROW tables).

with s (val_col_1, val_col_2) AS  (values 
            (231, 'Value 1'),
            (481, 'Value 2'),
            (513, 'Value 4')
        )
, i as (select * from final table(
    INSERT INTO ssz_me_mis.test_update_table ( val_col_1 , val_col_2) 
     select * from s where not exists (select 1 from ssz_me_mis.test_update_table t
        where 
        t.val_col_1 = s.val_col_1
        and t.val_col_2 = s.val_col_2
        )
))
, u as (select count(*) as dummy from new table(
    update ssz_me_mis.test_update_table t
    set val_col_1 = (select val_col_1 from s where t.val_col_1 = s.val_col_1 and t.val_col_2 = s.val_col_2)
    ,   val_col_2 = (select val_col_2 from s where t.val_col_1 = s.val_col_1 and t.val_col_2 = s.val_col_2)
    where exists    (select val_col_2 from s where t.val_col_1 = s.val_col_1 and t.val_col_2 = s.val_col_2)
))
select ID_col from i, u

I included a branch for updates, but logically you need some non-key columns for that to make sense. Your example is just an INSERT in practice, so I was a bit confused why you were using MERGE at all.