Robert Lichtenberger <r.lichtenberger@xxxxxxxxxxx> wrote: > I am trying to fully understand, how costs for queries are computed. > Taking the following example: > > CREATE TABLE test (name varchar(250) primary key) ; > INSERT INTO test (name) VALUES(generate_series(1, 1000)::text) ; > ANALYZE test ; > EXPLAIN SELECT * FROM test WHERE name = '4' ; > > I am getting the output: > Index Scan using test_pkey on test (cost=0.00..8.27 rows=1 width=3) > Index Cond: ((name)::text = '4'::text) > > The server has default cost parameters > > The value I want to understand is 8.27. From reading the book > "PostgreSQL 9.0 High Performance" I know, that we have one index page > read (random page read, cost=4.0) and one database row read (random page > read, cost=4.0) which comes up to a total of 8.0. But where are the > missing 0.27 from? > > If I modify the example to insert 10,000 rows, the cost stays the same. > Only if I go for 100,000 rows will the computed cost increase to 8.29. > > Can anybody enlighten me, please ;-). There are some other costs, in your case cpu_tuple_cost and cpu_index_tuple_cost. Andreas -- Really, I'm not out to destroy Microsoft. That will just be a completely unintentional side effect. (Linus Torvalds) "If I was god, I would recompile penguin with --enable-fly." (unknown) Kaufbach, Saxony, Germany, Europe. N 51.05082°, E 13.56889° -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
