Nsryjdtyk

I mean:

INSERT INTO test VALUES(1, 'message'), (2, 'message'), (3, 'message);

triggering will cause the result in the table to look like this:

1, E'messagenmessagenmessage'

How to forbid inserting rows and then continue operations on the transferred data in the insert?
I am using postresgql.

In Postgres 10+ you can use a transition table in an AFTER trigger, see Example 43.7. Auditing with Transition Tables. Assuming that id is a primary key (or unique):

create table my_table(id int primary key, message text);

you can update one and delete the remaining inserted rows:

create or replace function after_insert_on_my_table()

returns trigger language plpgsql as $$

declare r record;

begin

    select 

        array_agg(id) as ids, 

        array_to_string(array_agg(message), e'n') as message

    from new_table

    into r;



    update my_table

    set message = r.message

    where id = r.ids[1];



    delete from my_table

    where id = any(r.ids[2:]);



    return null;

end $$;

In a trigger definition declare a transition table (as new_table):

create trigger after_insert_on_my_table

after insert on my_table

referencing new table as new_table

for each statement 

execute procedure after_insert_on_my_table();

In earlier versions of Postgres you can simulate a transition table introduced in Postgres 10.

Test it in db<>fiddle.