Agree totally - the most flexible (and correct) way to do this would be model as a separate dimension. However if you absolutely cannot do that (for whatever reason) then you can build a logical column on this and simply subtract the two dates, convert to hours and "bucket" with a case statement. If you don't have RPD…

Not sure why you'd want to evaluate_aggr? This is generally used to allow you to use aggregate functions at the database level (i.e. not available in OBIEE) - so it pushes what you write in there down to the database level and applies a group by.

Ensure that you have defined unique keys at each level in your custom hierarchy.

Write-back in OBIEE is torturous at best and suicidal at worst. Not sure why it's even a feature because I've yet to meet anyone where it does everything they want it to as standard. Have you thought about having an APEX application that handles the writeback, and then call it from the OBIEE analytic with values passed in?…

The obvious answer would be to use a pivot view. However I assume that this is required for something more complex. Other than that you are into the realms of "manual" pivoting, which only works if you know how many cities you have. So you would have a case formula in each column: Case when "location"."location level 4" =…

They have different purposes. DV is designed for taking data from multiple sources and "mashing it up" to produce insights and identify patterns in data etc. It's more of an end-user type tool. Answers is more for building structured dashboards etc. You can use both. If you think of an organization, you would generally…

Have you checked your version (and settings) of IE are certified for use with your OBIEE version?

The load plan structure is held hierarchically in snp_lp_step, which joins to snp_loan_plan. I started something similar a long time ago in APEX but never got around to finishing it unfortunately. Monitoring Oracle Data Integrator (ODI) Load Plan Executions with BI Apps - Beyond Blog . It was always my intention to release…

If you have a single dimension which is shared between the two subject areas then it should have a common name.

Assuming you mean BI Apps customised files? Unfortunately it's not easy to identify them as other processes stamp the last update date and last updated by fields.

Also bear in mind that if you need to a calculation pre-aggregation then you must do that in the RPD using a physical expression. Physical expressions are evaluated pre-aggregation, logical expressions post-aggregation.

Did you monitor the session on the source database and verify it was actually running? You should be able to capture the SQL from v$sql and hence the plan. Then you can start looking at where the time is being taken through things like longops, active session history and so on.

Don't sort by a descriptive column - sort by your chronological key.

Generally lookups should be in dimensions and measures should be in facts. I wrote a couple of posts about combining facts with different levels of granularity. http://wegobeyond.co.uk/blog/entry/combining-facts-subject-areas-in-oracle-obiee.html

We've used APEX in the past, you just need to make sure your APEX application is resilient to URL injection attacks etc with it being hosted separately. On that last note, you'll need to configure your APEX HTTP listener to work cross domains if you are hosting APEX on a different server and want to use AJAX requests to…

I don't believe Act As is supported when using EBS authentication, primarily because you would need to go back and reinitialize your EBS session. I.e. you might have an initblock which does: Select organization_id from hr_organization_units. As that is a secure view, it is dependent on your EBS context (i.e. your user_id,…

Not that I'm aware of.

We similarly have hit this issue. Our client has an SR with Oracle but currently no resolution.

You can achieve things like this using presentation variables, however your requirement sounds "odd". You should try and do as much as possible in the OBIEE repository, and if that means having a logical formula that divides something by the distinct count of something else then do it that way.

Why do you need to use Evaluate to consider the previous row? I presume you mean previous as in chronologically previous? If so then that's what the "Ago" function is for. Do they wish to use the staging table because it only contains changed rows (i.e. incremental) rather than the full dataset?