Melkij

The sub-statements in WITH are executed concurrently with each other and with the main query. Therefore, when using data-modifying statements in WITH, the order in which the specified updates actually happen is unpredictable. All the statements are executed with the same snapshot (see Chapter 13), so they cannot "see" one another's effects on the target tables.

UPDATE "TABLE1"
SET
  "Value2" = (NOT EXISTS(
    SELECT NULL
    FROM "TABLE2"
    WHERE "что-то" = "кое-что"
)
      AND NOT EXISTS(
    SELECT NULL
    FROM "TABLE3"
    WHERE "что-то" = "кое-что"
))
WHERE "кое-что" = ANY ($1 :: INT [])

create unlogged table list_for_delete (id int);
insert into list_for_delete values ....

with to_rm as (
    select id from list_for_delete limit 10000
), rm as (
    delete from list_for_delete where id in (select id from to_rm)
)
delete from tablename  where id in (select id from to_rm);

vacuum analyze tablename;
drop table list_for_delete ;

Как по скорости и надежности такая схема по сравнению с просто RAID5 из всех дисков

3. Какой RAID предпочтительнее для виртуальных машин?

where array(select (j->>'category_id')::int from jsonb_array_elements(data->'items') j) && array[1,3,5];

?| text[] Do any of these array strings exist as top-level keys?

select * from t where jsonbfield ?| array['0', '5', '7']::text[];

select * from t where jsonbfield ?| (select array_agg(t) from json_array_elements_text('["0","5","7"]') t);