PostgreSQL, how to speed up checking if a column has a unique value?

Question

PostgreSQL

There is a simple table:

CREATE TABLE public.notices (
  id TEXT NOT NULL,
  json TEXT,
  CONSTRAINT notices_idx PRIMARY KEY(id)
) 
WITH (oids = false);

json may contain long text.
Before adding records, I check their presence in the table, this stage has gone through three stages of development:

1. One by one check:select * from notices where id = '123' limit 1);
   
2. One by one check:

   select exists(select 1 from notices where id = '123');

   
   
3. Checking the whole crowd:

   (SELECT id FROM notices WHERE id='123' LIMIT 1) UNION ALL (SELECT id FROM notices WHERE id='234' LIMIT 1) UNION ALL ...

   
   

The last method gave a significant speedup, but the number of records grows, the number of id in the request is large, so the speed of checking a pack of 50 thousand id for a database of ~ 1 million records is from 20 to 60 seconds on the current machine.
I would like to speed up the check to a minimum. How can you optimize?
Since the json field can be very long, I blame him too. If I allocate id to a separate table, is there any speedup?

Answer 1

There is an index for the id field, is it the primary key?
3 check - in my horror. Better like this:
SELECT id FROM notices WHERE id='123' OR id='234'...

Answer 2

You actually have access by key, store in key-value storage simply, for example in Redis.

Answer 3

Probably in most cases the objects are already there, so what's the best way to get the missing ones?

select "id" from "notices" where "id" not in ('1', '2', ...)