Question : left outer join not working

this is the query i am trying to build:
=====
select distinct grid.distrib, grid2.FiscYear, grid2.Quarter, grid2.Region, grid2.approved, count(grid2.Approved) as gridcount
from grid left outer join grid as grid2
      on grid.distrib = grid2.distrib
where grid2.FiscYear = 2006 and grid2.Quarter = 2 and grid2.region = '1'
group by grid.distrib,grid2.FiscYear,grid2.Quarter,grid2.Region,grid2.approved
order by grid.distrib
------------------
RTE             2006      2      1      0      1263
=====

there are 2 unique records in the distrib column that i would like to see whether there are matching records in the join or not.
=====
select distinct distrib
from grid
------------------
HSE
RTE
=====

same issue with this query.  i am only getting matching records like an inner join, but i am specifying a left outer.
=====
select distinct grid.distrib, grid2.FiscYear, grid2.Quarter, grid2.Region, grid2.approved /*, count(grid2.Approved) as gridcount */
from grid left outer join grid as grid2
      on grid.distrib = grid2.distrib
where grid2.FiscYear = 2006
------------------
HSE             2006      1      m      0
HSE             2006      2      m      0
HSE             2006      3      m      0
HSE             2006      4      m      0
RTE             2006      1      m      0
RTE             2006      2      1      0
RTE             2006      2      m      1
RTE             2006      3      m      0
RTE             2006      4      m      0
=====

it seems like the where clause is filtering on the left side of the join as well, even though i have specified the right side table for all the filters.  i am joining the tables on the distrib column because thats the only similarity i would like to have between the tables.  i have never had an issue like this calling from different tables.  is this because i am trying to join the same table to itself?

Answer : left outer join not working

grid.distrib is not unique
so the join brings back many more records than you expect hence your 1263 value
I'd be tempted to rewrite this completly but irst you could try
 count(distinct grid2.Approved) as gridcount

otherwise I'd go for

select distinct grid3.distrib, grid2.FiscYear, grid2.Quarter, grid2.Region, grid2.approved , count(grid2.Approved) as gridcount
from
(select distinct grid.distrib from grid) as grid3

left  join (select * from grid where FiscYear = 2006 and Quarter = 2 and region = '1') as grid2
     on grid3.distrib = grid2.distrib
group by grid3.distrib,grid2.FiscYear,grid2.Quarter,grid2.Region,grid2.approved
order by grid3.distrib

I think that should cut down your duplicates.

any help?