If this approach isn't working for you (not sure why), you could try using the lookup dense/sparse as a way to achieve the same results. For every sale look up the cost based on the keys ... Oracle BI EE 11g – Sparse Lookups & Left Outer Joins - Rittman Mead Consulting

The easiest way to get the report column and the prompt column the same is use the same presentation column -- you're best served by doing the case in a logical column and putting it in your subject area (sometimes I will create a sub-folder special for prompting and filtering for these types of columns) ...

@rmoff wrote this up ... http://www.rittmanmead.com/2012/09/automated-monitoring-of-obiee-in-the-enterprise-an-overview/

You're welcome ... it's how I do it, can't remember if I learned the hard way or not! 

Your issue is more of a data model matter versus a tool matter ... +1 to rmoff's and Mr. Berg's comments

Yep ... I mistook it ... the values are what you want to shorten ... good catch!

what about simply copying the column that you are using and changing the header either in the formula or the column properties window?

I think the OP is asking how to process source deletes in the warehouse ... so they are looking to implement a mechanism.

in the following example I had month (YYYY/MM) and revenue ... I added another column (Month || ' in ' || Quarter) ... then used that as my horizontal axis column ... just make sure you create your custom column at the same grain as the x-axis column you really wanted to use (in your case looks like day or date)

I highly advise avoiding an actual delete from the warehouse ... use soft-delete.

It is in the documentation ... http://docs.oracle.com/cd/E28280_01/bi.1111/e25318/getstarted.htm#BIEMU11338

check out that blog post before you go after the driving table ... Driving tables are for cross-database joins ...

Try running it from the command line and have the OBIApps RPD and the output (trimmed) RPD in the same folder as the script ...

This might help you: https://obibb.wordpress.com/2010/05/31/multiple-fact-reporting-on-non-conforming-dimensions/ Approach is the same despite this being a 10g related blog post.

Use the Physical layer joins to specify columns Use the logical layer joins to specify inner/outer, cardinality, driving table, etc.

I did see the original write up ... what I'm proposing to eliminate the logical step (in the BI server) AND there's no injecting of thousands of ID into a where clause ... Nor does this invoke any new ad groups - I merely suggest you could use them to be able to keep the security table current.

Your layout won't work as your Variance % is not a dimensional attribute of Time - it's a measure label. The force of a good BI program is: Answers at the speed of thought - can people get accurate information and does the system perform? Delivery at the speed of business - can new insights be developed and made available…

* Data Exploration & Discovery - use to uncover questions about your business you didn't know you needed to ask * Analysis and Interactive Reporting - use to interactively answer known questions about your business; built on an information model designed knowing the use of information; on-screen decision making * and…

calculate your variance in the criteria tab, then put measure labels as the row and month-year as the columns ...

1) You can send in direct SQL - BUT you circumvent the investment made in the BI stack from Oracle AND you have limited functionality with this approach. Not recommended. 2) A subject Area is a representation of the logical (business rules and relationships) model which is based on the physical data model. So no, just…