Re: Request for some tips or better ways to handle some "issues"
From: Tom Chester (publicNOSPAM_at_tomchester.net)
Date: 03/09/04
- Next message: Mosha Pasumansky [MS]: "Re: Writeback temporarily to a cube with a Excel pivot table"
- Previous message: Tom Chester: "Re: Filtering Dimensions"
- In reply to: Jim Burnham: "Request for some tips or better ways to handle some "issues""
- Next in thread: Jim Burnham: "Re: Request for some tips or better ways to handle some "issues""
- Reply: Jim Burnham: "Re: Request for some tips or better ways to handle some "issues""
- Messages sorted by: [ date ] [ thread ]
Date: Tue, 9 Mar 2004 11:00:02 -0800
See "TC" replies interspersed below.
public @ the domain below
www.tomchester.net
"Jim Burnham" <jim_burnham_home@yahoo.com> wrote in message
news:E75937DB-B23F-4D20-82DB-2E81B2EC0005@microsoft.com...
> Hi. I have been working to recreate an existing Essbase application in
Analysis Services as a learning exercise. I have a few things which took me
a little time to figure out and was wondering if anyone had suggestions on
better ways to do them or tips for me. I am not knocking the product or
making a big deal out of little issues, just looking for a better way to do
some of the things I tried to do.
>
> Thanks in advance for any suggestions,
>
> Jim
>
> ****************************
> I like the ability to modify a shared dimension without changing the
already constructed cubes, but am not quite clear on when the cubes get the
changes I make to the dimensions. When I finish modifying the dimensions
and process them, what does that do? I still can't browse the dimension
until at least one cube gets updated. Am I correct in thinking that the
processed cubes will continue to function as is after modifying a dimension,
until I re-process them?
TC: It depends on what the change is. When you make a structural change to a
shared dim, cubes that contain that dim go offline.
>
> I frequently want to filter the dimension members loaded into a cube to
only include ones for which that cube will have data. I find myself
sourcing the dimensions from SQL views which join back to the fact to ensure
this. Is that a good way to do this? I have also made the mistake of not
including SELECT DISTINCT in those views which really wreaks havoc.
TC: Yes, I believe in always using views.
>
> I have never used the "member key" settings - I just click "next". Should
I be using the dimension members' surrogate keys or something other than the
members' names?
TC: Yes.
> The storage design wizard is very easy to work with, but is there also the
option to get under the covers? If I had a dataset where I knew I had
better data density across a particular combination of dimensions, or do
more frequent analysis across a particular dimension, can I design my
storage for that?
TC: You can get under the covers (to an extent) with usage based
aggregation, or by writing DSO code. Also be aware of the aggregation usage
properties (dim and level).
> Is there an easy way to build grouping members on the fly to make
navigation easier?
TC: No.
>
> I have noticed that when I use views as source for the dimensions, and
change the underlying view columns, I have to actually exit Analysis
Manager. Is there a way to get it to refresh without exiting?
TC: Right-click the data source and choose refresh.
- Next message: Mosha Pasumansky [MS]: "Re: Writeback temporarily to a cube with a Excel pivot table"
- Previous message: Tom Chester: "Re: Filtering Dimensions"
- In reply to: Jim Burnham: "Request for some tips or better ways to handle some "issues""
- Next in thread: Jim Burnham: "Re: Request for some tips or better ways to handle some "issues""
- Reply: Jim Burnham: "Re: Request for some tips or better ways to handle some "issues""
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
