Tabular: fast patterns for dynamic banding of measures by 'effective date'?

Question

I've implemented
http://www.powerpivotpro.com/2012/04/aged-debtorsdynamic-banding-in-dax/ but it's very slow - 5min+ to render a simple excel pivot table that's filtered to 2 invoices from a model that's ~130meg and has around 500k invoice payment
schedule rows in the largest fact table. Cube deployed to server as opposed to power pivot/excel. I changed some of the logic to get rid of the 'as at' dynamic filtering in the linked blog and use a timeline instead to select all the dates from
the beginning of time to my effective date, but it's still 30sec+ refreshes.. Just wondering if there's better patterns out there that use v2 dax and are faster. My transposing might've also introduced some inefficient logic because I don't have a central
'transactional' table to work from.
My model looks like - [table](relevant columns,...)
[invoice](invoice date) 1:m [payment schedule](due date, amount due) 1:m [Payment] (payment amt, payment date)
in words: an invoice is raised. An invoice can be made up of part payments that are due at different times, a payment can be for part or all of a payment schedule.
The model works for 'as at now' reporting, but I'd like to implement a banded aged debtor report as at a specific date in a way that is usable.
My current implementation doesn't include all the features and it's already pushing a minute of excel processing for a very filtered subset of data. Don't want to imagine how long it'd take for the full set.
some data
-- invoices
SELECT *
FROM (Values
(1,'1 jan 2015'),
(2,'15 jan 2015'),
(3,'20 jan 2015'),
(4,'21 jan 2015')
) as t(InvoiceID,InvoiceDate)
--payment schedule
SELECT *
FROM (Values
(1,1,'1 feb 2015',$100),
(2,2,'15 feb 2015',$50),
(3,2,'15 mar 2015',$50),
(4,3,'20 feb 2015',$50),
(5,3,'20 mar 2015',$50),
(6,4,'21 feb 2015',$25),
(7,4,'21 mar 2015',$25),
(8,4,'21 apr 2015',$25),
(9,4,'21 may 2015',$25)
) as t(PaymentScheduleID,InvoiceID,DateDue,AmountDue)
-- payments
SELECT *
FROM (Values
(1,1,'10 Jan 2015',$50),
(2,1,'11 Jan 2015',$50),
(3,2,'10 feb 2015',$50),
(4,4,'25 feb 2015',$50)
) as t(PaymentID,PaymentScheduleID,DatePaid,AmountPaid)
The bands I have are
BandID Band Label BandStart BandEnd
1 Future -999999999 -31
2 Current -30 0
3 1-30 days 1 30
4 31-60 days 31 60
5 61-90 days 61 90
6 91-120 days 91 120
7 120+ days 121 999999999
anything due more than 30 days into the future is 'future', anything due in the next 30 days is current, anything that's due in the past is overdue.
The 'effective date' filter should exclude any invoices that were raised in after the effective date and exclude any payments made after the effective date.
expected output:
Thanks
Jakub @ Adelaide, Australia Blog

Answer 1

See if the blog below helps:
https://cwebbi.wordpress.com/category/dax