True, unfortunately sometimes you just have to model 1:1 what you are given, no matter is it wise or not. :D

Thanks, I'll test which one works best :)

So I'd need a join between contact and reservation and looks like I really need to multiply probably the reservation, which is pretty stupid thing in end users point of view.

Attached the template (second table in it) and xml data. BR, Mari

I read the documentation but didn't understand how should I modify this to get it work. Do you have any tips how to do it easily? BR, Mari

Solved this by calculating everything in the database. Basically I created the year sum on January on a new table column, imported the column to RPD, changed default aggregate to sum and then created a new OBI column on top of that. The new OBI column is normal cumulative calculation; todate(myNewDBColumn, year) and that…

That's the method I use too, put the process in pieces and test piece by piece. Usually people just get blind to their own work; I'm pretty sure that this is some tiny minor thing somewhere but I just can't find it! DB is ok, I'm 100% sure of it. Physical layer is ok, I get everything I expect to get. Base BMM is ok…

No matter if I use base measure (physical table column) or formula, result is wrong (tried both). Formula is normal todate(measure, level) aka todate(mymeasure, year) in this case.

Yes and I did that duplication/rename/pin - unfortunately it multiplies the data as I described. No idea why, duplication works perfectly in HR models and they are built exactly the same way as this one.

Yes, no luck - of course 53 million euros looks nicer than 600 000. I tried this also so that I created a new view to db that calculates the year totals and joined that as another fact to my subject area. It works but it would cause major changes to my BMM if I finalized it properly (needs additional dimensions that I…

It's dummy enough, thanks. But still it won't work. If I filter like that and prompt for 201706 (or anything except 201712), the filtered column shows empty. Basically it's logical; if prompt filters the data to a certain month, I can't expect it to show data for another month at the same time. This has been one %&/!" last…

In RPD everything else works just fine except the whole year calculation: - base measure ok - cumulative base ok - base -1 year ok - cumulative base -1 year ok - all differences and percentages between them ok - cumulative base -1 year for year level does not work and no calculation in Answers works either

But how do I do it? I need specifications for dummies. I tried FILTER("Cumulative EUR last year" USING "D1 Time"."Month"='201712') and ok, I get what I want to that certain column BUT my prompted months won't work after that and if my selected month is anything else but that 201712, the column will be empty. If I could…

"sum 'Eur This Month' (full syntax, don't know your folder names!) group by Year (ditto) as a formula should work." Yes it should. Really. But it won't and that's the problem. I have never understood how that filter syntax works with dynamical values, it could probably solve this. "Did you notice that the amounts are equal…

I don't use filters unless I filter the whole analysis. Column headings are done like this, all in Admin: Cumulative EUR this Year: TODATE(Eur This Month, Year). Aggregation level is month so it calculates sum(jan:this_date) Eur This Month: Base fact column, no calculations. Aggregation default (sum) Cumulative EUR last…

This is a very simple mode with one fact and one dimension, levels are correctly set. Could you give me a working example how to do this in answers? What I got from Oracle support doesn't work anymore, neither anything I have tried myself. And no, there's no case clauses in the analysis at the moment. I start to suspect…

1-3, Yes. I have checked all joins, all hierarchy keys and tried server complex aggregate and no, still quite interesting amounts. I've even tried to calculate this in db but it's a bit heavy way to do this.

I mean sum(Jan:Dec) and I have done exactly that but it doesn't work, no idea why. Here you see what happens: