Re: tricky GROUP BY / JOIN question

T E Schmitz <mailreg@numerixtechnology.de> writes:
> This is *almost* what I need:

> SELECT
> BRAND.BRAND_NAME,
> MODEL.MODEL_NAME,
> min (ITEM.PRICE),max (ITEM.PRICE)
> *min (CONDITION.POSITION),max (CONDITION.POSITION)*
> FROM ITEM
> left outer join MODEL  on MODEL_PK =ITEM.MODEL_FK
> left outer join BRAND  on BRAND.BRAND_PK =MODEL.BRAND_FK
> left outer join CONDITION on CONDITION.CONDITION_PK = ITEM.CONDITION_FK
> group by BRAND.BRAND_NAME,MODEL.MODEL_NAME

> In the result I don't want min/max(POSITION) but CONDITION.NAME of min 
> and max(POSITION) for each MODEL.

I think you could do something like

SELECT
BRAND.BRAND_NAME,
MODEL.MODEL_NAME,
min (ITEM.PRICE),max (ITEM.PRICE)
(select name from condition c1 where position = min(condition.position)),
(select name from condition c2 where position = max(condition.position)),
FROM ITEM
left outer join MODEL  on MODEL_PK =ITEM.MODEL_FK
left outer join BRAND  on BRAND.BRAND_PK =MODEL.BRAND_FK
left outer join CONDITION on CONDITION.CONDITION_PK = ITEM.CONDITION_FK
group by BRAND.BRAND_NAME,MODEL.MODEL_NAME

ie do a sub-select to get the desired name.

You need Postgres 7.4 or later to get this to work --- before that we
would have mis-interpreted the aggregate calls to indicate aggregation
within the sub-selects.  The current interpretation is per SQL spec:
since the aggregate argument is a variable of the outer select, the
aggregation occurs with respect to that select, and the aggregate result
is passed down to the sub-select as a scalar.
        regards, tom lane