Just a follow-up, the suggested work-around works great. I had to move the
column_properties outside the class definition for some of them to work,
but the final result works fine (classes not shown but the definitions are
pretty obvious):
def IF(condition, true, false):
""" Construct a simple case expression """
return case([(condition, true)], else_ = false)
O = Order
L = OrderLine
P = OrderPayment
N = Numeric(14,2)
N.get_col_spec = lambda: "NUMERIC(14,2)"
def total(taxable):
""" Helper to define total column properties """
expr = L.quantity * L.price
if taxable is not None:
expr = IF(taxable == L.taxable, expr, 0)
return func.sum(cast(expr, N))
O.taxable = column_property(
select([total(True)], L.order_id == O.id).label("taxable"))
O.nontax = column_property(
select([total(False)], L.order_id == O.id).label("nontax"))
O.subtotal = column_property(
select([total(None)], L.order_id == O.id).label("subtotal"))
O.sales_tax = column_property(
cast(func.round(O.taxable * O.tax_rate / 100, 2), N).label("sales_tax"))
O.total = column_property(
select([O.subtotal + O.sales_tax]).label("total"))
O.discount = column_property(
cast(0.00, N).label("discount"))
O.payments = column_property(
select([func.sum(P.amount)], P.order_id == O.id).label("payments"))
O.change_due = column_property(
select([O.payments - O.total]).label("change_due"))
del O, L, P, N, total
Thanks once again for your assistance.
Shawn
On Sun, Mar 22, 2009 at 1:14 PM, Shawn Church <[email protected]>wrote:
>
>
> On Sun, Mar 22, 2009 at 12:39 PM, Michael Bayer
> <[email protected]>wrote:
>
>>
>>
>> On Mar 22, 2009, at 3:00 PM, Shawn Church wrote:
>>
>> >
>> > This isn't really the same thing because the float would be rounded
>> > after the sum not line by line. I think I could also use sum(...
>> > type_ = Numeric(14,2)) for the same result. Am I missing an easier
>> > way to do this?
>>
>> its an SQLA bug that will require a workaround until 0.6. another
>> approach is monkeypatching a get_col_spec() on your Numeric type:
>>
>> n = Numeric(14, 2)
>> n.get_col_spec = lambda: "NUMERIC(14, 2)"
>>
>> that will work throughout 0.5. in 0.6 it will not be needed but also
>> have no effect.
>>
>>
> Thanks, this seems a reasonable workaround. But would I cast(..., n)?
>
>
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~----------~----~----~----~------~----~------~--~---
