Hi all.
We got bit of an issue with linked partitioning.
Let me give a brief you the requirement, that way you would know if we are doing something wrong.
The client wants to report on the transactional level. For example, a particular account has many transactional information, some are numbers such as transaction id and some are textual descriptions. In total there are almost 20 to 30 of them that define a transaction.
Initially we tried to load all the transactional information into essbase. But the problem is with text measures. Our understanding is that Essbase only supports 1024 entries of text measures, whereas we have much more data than that.
Then we decided to keep all the high level informations in a BSO cube and all the transactional level information in an ASO cube. The two cubes are identical except the fact that ASO has all those transactional informations as dimensions. So, now, for example, Transaction ID is not a measure any more, it's a dimension which has probably 3000 children.
We planned to create a linked partition between the ASO and BSO cubes so that when we do Smart View ad-hoc, we do it on a higher level of BSO, and only when we need, we drill through to ASO for transactional information.
Are we doing the right thing? Is there any other way of achieving what we're trying to achieve? We don't have access to the source system, we are given only text csv files, so, Essbase Studio is not an option, right?
We tried to do the Smart View, it gives us some special character in text error. Our version is 11.1.2.3. As this version does not support Escel Add-in (well, that's what we found on the Internet), we wanted to replicate the process in 11.1.1.3 with Excel Add-in.
Now, when we do an ad hoc with BSO and drill down to ASO, whatever extra dimensions (all the transactional informations as mentioned above) it has, they all appear with higher level. And it looks like the user has to change POV probably thirty times as there are 30 dimensions.
How does a linked partition work? What is the best approach when there are thousands of textual measures to report on?
Any kind of idea, suggestions would be highly appreciated.
Thanks