At 16:26 04/05/2007, you wrote:
Paul Smith wrote:
Why doesn't it use the other index? If use 'set enable_seqscan=0'
then it does.
Just a guess, but is the table clustered on column a? Maybe not
explicitly, but was it loaded from data that was sorted by a?
I wouldn't ha
This is in Postgres 8.1.5
I have a table like
CREATE TABLE x (a VARCHAR, b VARCHAR, c VARCHAR);
CREATE INDEX y on x(a);
CREATE INDEX z on x(b);
There are over a million rows in 'x'. Neither a nor b are unique.
There are probably about 20 or so distinct values of a and 30 or so
distinct values
