Are you getting data from multiple dimensions? If so, I'd suggest gradually removing dimension attributes until you pin point which dimension has the problem and then you can look at your logical joins.

The joins that OBIEE creates are based on your rpd model. What does your model look like between your logical dimension and your logical fact?

Install OBIEE on a server.

I suggest you take a look at Oracle's certification matrix where you'll see that Windows 10 isn't a supported OS. Also, search through this forum where you'll see that several others have experienced similar issues by installing enterprise software on a desktop OS.

Windows 10 is a desktop OS and OBIEE runs on a Server OS. You'll need to reinstall on a Server that has a certified OS. You just say you have OBIEE 12C and there are a number of versions that all have different certified OSs.

What query do you see generated by the BI server in the log files? Are there any exceptions in the log files?

Have you been able to connect directly to the database as the reports user in SQL developer? If so we’re you able to select from these tables?

If you have a date dimension then you could add the fiscal month period as an attribute to the date dimension and then you can very easily derive the period each date belongs to.

By 1-12, do you mean month numbers?

Without even checking, I can tell you that it won’t be supported because the OS and DB are newer than OBIEE 11g. I would strongly suggest that you move away from 11g which is very old to say the least. if you need 11g, then you’d need a dedicated server with an older OS that it is certified to run on. As you’re aware,…

You could the search/replace functionality in catalog manager to search for the logic and replace it with the 12C fix. - https://docs.oracle.com/middleware/bi12214/biee/BIESG/GUID-FF29B794-97F5-4975-A924-ED40319B9150.htm#BIESG334

Aaccording to the 12.2.1.4 certification matrix, OBIEE q2.2.1.4 is certified to work with Windows 2016, SQL Server 2016.

The Cache results are based on actual data from the underlying data source. The "NULLS" don't exist in the data and the BI Server does the manipulation after the data is returned and before it is displayed in your analyses.The only way to get the "NULLS" cached is to store the NULLS as facts in your database.

Have you checked the Session Manager in your Admin Tool: Manage > Sessions where you can see the value of all variables per user logged on?

Have you by any chance considered theData Visualisation tool? This is much more self service focused and may be a better fit for your use case.

Did you see this document from Oracle?

Did you see this document from Oracle?

A few things you could try: You can take a snapshot of 1 analysis that has a date column that is affected by this. Make the changes to the date column custom format in this analysis and then compare the XML of the analysis before and after the change to see what the XML changes are. You can then make the global change by…

Take a look at this document from MOS

https://docs.oracle.com/middleware/12211/biee/BIESG/GUID-464861F6-CD14-443F-9251-E1C5300CEB02.htm#BIESG1674