I have an issue with a dimension that uses customer data and getting the
64,000 children issue.

Before I explain I have looked at the kbase article:

INF: How to Work with More Than 64,000 Children Per Parent WGID:223
ID: 303969.KB.EN-US
http://support.microsoft.com/default...b;en-us;303969

And attempted to implement, but I'm still getting the issue 64K error message.

Working on forecasting churn and customer triends were going against a table
from the data warehouse that contains the historical information about each
customer. This table has multiple entries for each customer and has a unique
id on each record.

The data is logically arranged from the national level down to the customer
data as follows in a shared dimension:

Region (5 regions)
Channel (12 Channels across the 5 regions)
Area (57 areas across the 12 channels)
Account (approx 1.5 mil)

When I implement automatic groups using the above hierarchy and specify the
Unique ID field as the key column It fails on the Channel stating that there
are more than 64,000 children under it. Yes there are more than 64K accounts
to a channel but that should get rolled into Area and then into Channel.

Here is what the dimension actually looks like

Region
Channel Group
Channel
Area Group
Area
Account Group
Account

Should I flip this hierarchy so the Account is on top? Implement a hidden
group between Region and Channel that splits the customers along alphabetical
lines?