Fast alternative to COUNT(*)?

Question

PostgreSQL

Now I use SELECT COUNT(*) СложныйЗапрос
COUNT (*) predictably slows down. How to speed it up without big edits?

Arrange options:
1. Return an approximate amount
2. Return COUNT (*) for the first 1000 elements. They say that there are so many files falling under the condition in the database

SELECT COUNT(*) FROM files t1

    WHERE NOT EXISTS (SELECT 1
              FROM files t2
              WHERE t1."Number" = t2."Number" AND
                t1."docDate" < t2."docDate"
            )
    AND t1.status IS NULL
   AND t1.section_name='Users' AND t1.status IS NULL

The subquery is needed to select the newest version of the document. there may be duplicate numbers in the database

Answer 1

It was yesterday!
How to quickly count the number of records in a table with a billion records without an index?