[curiousgally]

But in my scenario where 4 columns need to form a hierarchical facet.it's painful to do so many if's and concatenations. ... :)

[Isomorphic]

Guess what? It's in the QuickStart Guide sections on DMI, specifically under Adding DMI Business Logic. A DMI is called before the query is executed, dsRequest.execute() executes the query and the DSResponse contains the results, which you can modify.

It's really worth reading every single sentence of the QuickStart Guide, it is packed with key information.

[curiousgally]

yep, reliased it so edited the post....how lame a question i had..??

[curiousgally]

From my observation for a cube grid.

In a tree facet collapsing of levels can be configured. Depending on this configuration cube Grid requests for data firing a query. So if the data is huge then does it do paging like displaying as and when it fetches or does it wait till the whole data for that collapse is available.

[Isomorphic]

At any given time the CubeGrid only fetches data needed for display.

So if you provide a tree facet that is initially partly collapsed, data is not loaded for nodes that are not shown. Then when the tree is expanded, data is loaded for the revealed nodes.

Data is also only loaded for visible cells. If you have a large cube showing scrollbars, data is loaded as you scroll horizontally or vertically and reveal more cells.

[Koripella]

1) in the sample, it's pre-calculated in the dataset. You can either pre-calculate it as well or calculate it on the fly on the server. It can't be calculated client-side since load on demand means only part of the data is available.

2) you can modify dsRequest.criteria server-side to split the list of ORIG values between your two columns.

I have a question regarding the above two quoted cases.
Can you please provide a sample code for
- calculating the sum (if possible both the ways)
- modify dsRequest.criteria server-side to split the list of ORIG values between your two columns