sql - Understanding finding max value with index in postgreSQL -

i have following table:

create table foo (   c1 integer,   c2 text ) 

created follows:

create indedx on foo();  insert foo   select i, md5(random()::text)   generate_series(1, 1000000) i; analyze baz; 

now, tried query:

explain (analyze, buffers) select max(c2) foo; 

and got following plan:

result  (cost=0.08..0.09 rows=1 width=0) (actual time=0.574..0.574 rows=1 loops=1)   buffers: shared read=5   initplan 1 (returns $0)     ->  limit  (cost=0.00..0.08 rows=1 width=33) (actual time=0.570..0.571 rows=1 loops=1)           buffers: shared read=5           ->  index scan backward using foo_c2_idx on foo  (cost=0.00..79676.27 rows=1000000 width=33) (actual time=0.569..0.569 rows=1 loops=1)                 index cond: (c2 not null)                 heap fetches: 1                 buffers: shared read=5 

what confused resulting cost 0.08.. 0.09. why?

i thought find max, if had index on column had perform index scan , read @ least 1 of index leafs. reading leafs in turn accomplished 1 random acces costs 4. so, cost should have been more 4.

what did miss here?

the cost of index scan pro-rated limit. proration logic not try take rounding of page accesses integers account, @ point proration done has been collapsed single floating point number.