Building a Postgresql View Over a View - sql

Question

I've got a problem with one of my clients for OpenERP. I need to build a custom object for them, cause they have a problem where sometimes they get returned products after they have been sold. So I need to mark that as negative and subtract from the total sold from before.
SELECT sum(rsm1.product_qty_out) as "out"
, sum(rsm2.product_qty_in) as "in"
, rsm2.location_id
, rsm2.product_id
, rsm2.month
from report_stock_move as rsm1, report_stock_move as rsm2
where rsm2.location_id in (
select id
from stock_location
where "name" = 'Consumers'
)
and rsm1.location_id not in (
select id
from stock_location
where "name" = 'Consumers'
)
and rsm1.location_dest_id = rsm2.location_id
and rsm2.location_dest_id = rsm1.location_id
and rsm1.state = 'done' -- both are in done state
and rsm2.state = rsm1.state
group by rsm2.location_id, rsm2.month, rsm2.product_id
;
Anyone have any ideas? Also, I keep getting different results depending upon which table I use for the grouping. Why is that?
EDIT
Sorry, in my haste to solve this, I seem to have made my point clear. report_stock_moves that have a location_id of Consumers from a stock_location table need to be negative and summed up into report_stock_moves that have location_id of another place, but Consumer is the location_dest_id.
Example:
location_id = Some Place (actually ID of stock_location of course)
product_qty = 8
location_dest_id Consumers
date = 2012-07-02
location_id = Consumer
product_qty = 4
location_dest_id Some Place
date = 2012-07-04
Here are two record examples. They have different dates attached to them, so I'd probably want to take the max date, or the date of the returns(from Consumer to Some Place) if I want to union or group them so that when they are put together, i get 4, instead of say 8 if I didn't put them together and only look at what is going into Consumers.

Answer 1

In case anyone struggles with this for later. I'll paste my answer here. If anyone can help me explain about why union all in here seems to work, I'm all ears. I can do simple SQL, but I don't know much about how unions work, especially in case like this.
select max(total.id) as id,
sum(total.product_qty) as product_qty,
total.location_id as location_id,
total.date as "date",
total.year as year,
total.month as month,
total.day as day,
total.product_id as product_id
from (
-- From Consumer back to location
select -max(id) as id,
-sum(product_qty) as product_qty,
location_dest_id as location_id,
"date", year, month, day, product_id
from report_stock_move
where location_id in (
select id
from stock_location
where "name" = 'Consumers'
)
and "state" = 'done'
group by location_dest_id, "date", year, month, day, product_id
union all
-- From Location to Consumer
select max(id) as id,
sum(product_qty) as product_qty,
location_id, "date", year, month, day, product_id
from report_stock_move
where location_id not in (
select id
from stock_location
where "name" = 'Consumers'
)
and "state" = 'done'
and location_dest_id in (
select id
from stock_location
where "name" = 'Consumers'
)
group by location_id, "date", year, month, day, product_id
) as total
group by location_id, "date", year, month, day, product_id