Hi,
In Oracle there are instances when as a developer I know how the data
is spread in the tables and I want to be sure the database uses the
right index. Does the following hold true in Postgresql for a query
like this:
select s.order_id
from small_orders_table s, orders o
where s.find_these_id in
(select some_id from some_table where some_indexed_field = 'ABC')
and s.order_id+0 = o.order_id
and date_trunc('microseconds', o.valid_until_dt) < now()
This should essentially use the s.some_indexed_field as the primary
index and hit the orders table on the order id.
The +0 and date_trunc is used purely to ensure the valid_unit_dt field
(which is indexed) is not used.
Is this efficient?
Does date_trunc render the index invalid or can I do something else
(+0) doesnt work.
thanks
Tim
