Thanks I am a veteran of the instanceconfig file and upper limits on pivots, rows etc, but this does not answer my underlying question, why does an excluded column in a pivot table cause the error, I had understood that for a pivot table the data is retrieved from the server first and then pivoted on the web, so surely if…

Configure your prompt to populate a presentation variable. Reference the variable in your title using the @{My_Variable}{MyDefaultValue} Hope this helps?

Have you checked that the order execution precedence is still what it should be, so your NQ_SESSION.location is actually populated first?

Hi, web is not really my thing, but from the error it looks like it is preventing you building the body of the http, but is okay with you providing a query token along the lines of &my_token_to_pass=@{1} This would protect your other system from an informed destructive http injection attack.

If anyone lands on the page then the report will run, and it will be recorded, so you have to have a wholly unvisited dashboard tab for things not to appear.

Also meant to add, I have constructed a query with ALL other fact / dimension combinations and this is all 100% okay, the only point at which my error manifests is when I add the degenerate dimension in.

If you have usage tracking enabled and you are okay with the caveat of (and have been used in the last X days) then you could query it from that source.

Ok. I understand, but I am still left with my problem, and having checked fact / degen fact I don't see anything wrong, and I am puzzled as to why I get the error when I introduce the dimension hierarchy. I have tried deleting and starting over to double check everything and I still get the same issue. Any reason not to…

Thanks Christian, your last comment may be my problem, I had used the SAME alias for the source of both the FACT and its degenerate dimension. I had it in my head that as they were the same physical source then the logical join between fact and degenerate fact would resolve to; - select * from fact But what it was trying…

Hi, thanks for taking time, I know this should be easy, it just isn't.... Scenario 1: Fact table - joins to various dimensions, source 1 physical table (aliased). Dim Table - based on same source as fact table, logically joined. No hierarchy created, no content level as a consequence can be set. Result: No error in…

Hi, I have used the formulas in the past to build up a dynamic URL to drill out of OBIEE to another system, would this work for your requirement? If not, can you be a little less vague with what your error is, and also more fully explain where the parameter is getting its value?

You could use a row level variable populated via an init_block to place a pseudo security layer over your report, not what it is meant for, but it would achieve the desired effect.

Hi Christian, outer only works so far though surely? If A - B - C - D A to B outer B to C outer C to D outer Will we still get data in A if you include the entire chain; A,B,C,D; in the queries and the 'break' is in B? thanks, Robert.

Many thanks Christian, this was my suspicion!

lol, is that the party line, we gave you what you asked for, you never asked for it to work?! Thanks for ending my week on a belly laugh.

Hi, if the last day is or can be flagged explicitly, or can be identified by logic in some way, then you could apply a filter to the dimension in question to only display the last day of each month. You could also create a copy of your existing measure and put some case logic on that to make the value zero unless the month…

Thanks for this, this does better than some versions of the database, where what you describe will still not return data....

Sorry, reading your answer again you mean snowflaking on the LTS source, so you see it as if it is a single table, yes, and set the inner table to be the driving table - this is obviously NOT a no no, but I still think my original thoughts on the physical SQL would be true and it would not resolve my requirement? thanks…

Hi Christian, Snowflake in the business model is a big "no no", surely? And even then I am not getting how this solves the outer join and filter in combination issue? Surely the physical sql generated would then be; - select t1.VarFields, t2.VarField from t1, t2 where t1.ColX = t2.ColX(+) and t2.ColY = 'FlyInTheOintment'…

Thanks, I had thought of this but I am trying to avoid any kind of DB build if possible, I should have included MVs in my 'no go' list.