Btw - advising us up front that this was OTBI would have yielded better advice sooner, this is why in the how to ask questions they ask you to put your version of the software on the post.

If the AND logic still applied, and putting explicit filters on does not return any rows then SonPat99's suggestion will not work either - your logic is flawed. Removing the external filters and using the filter(filter logic should - provided there is data that satisfies both conditions you are putting on the aggregated…

See also; - So here you will substitute your quarter equivalent in your time hierarchy where the below has time.time.year, the 1 is for the number of quarters you want to go back, the first part is the measure you are looking back to the previous time based unit on. As Gianni / Christian have pointed out if your time…

On didn't get me results, did you remove your external filters?

Thanks for aiding on the understanding front Gianni - I have a feeling @3612030 (have you thought of making the display name more user friendly, it can still be anonymous?) is looking for the formula filter functionality that I briefly alluded to, to get KPI type functionality in a measure without the row by row situation…

And did you try my test with two explicit filters to ensure that the data in combination (in the same row) actually does exist?

On my last question, can you prove this for me; - "But can I repeat my question. On a row of data are you certain that both parts of the AND statement can be true at the same time?" Create explicit filters that would create the equivalent of what you are trying to achieve with the date based filter - see if it gives you…

But can I repeat my question. On a row of data are you certain that both parts of the AND statement can be true at the same time?

AND???? Are you sure the two are true in combination??

Try changing your filter to SQL and; - cast(replace(replace("Pipeline Snapshot Date"."Pipeline Snapshot Enterprise Quarter","Q","")," ","") as integer) = (cast(replace(replace('@{QUARTER}',"Q2","")," ","") as integer) -1) Cold coded - may need minor syntax fixes...

Probably also worth checking that your 2017 Q 4 is exactly what you see, check the length, just in case there is something you cannot see there. You might find it easier to strip the text and spaces out of it, then you can just take it away without all that cast and case.

Okay, if you have confirmed that your function returns exactly the same as your Quarters then I would change your IN to an equal sign; =

Amen to Joel's comment. But on your question - you seem to be missing a right parenthesis; - "Pipeline Snapshot Date"."Pipeline Snapshot Enterprise Quarter" IN (....................... ) END - there should be a bracket after that end statement. Also, in your example you show no space between Q and the number; - 2018 Q1;…

Agreed - I was just not certain what the precise meaning of clone was in the context - I agree totally with you, if that is what he meant!

'clone' was too vague for me to comment, I was not sure if we were in the physical or in the logical, and if by clone he meant alias or copy. Stylistically I like to alias ALL tables, as for me this prevents the 'is this an alias or a table question, they are all aliases, and I use the naming to give a sound steer on the…

Agreed on the storage, and agreed you would not want to try it on a massive star schema, but denormalisation is a good strategy to improve performance, joins do slow performance.

Call me a pedant, but can we make that "Attributes in physical facts are only there for RPD-level calculations and NEVER to be exposed except via logical Dimensions" - nothing will out perform a single table that is used as a star schema....

Take a look at this in my Oracle support; - After Upgrade To 11.1.1.9 RPD Consistency Check Fails With [15048] Logical Table: Field Must Be A Descendant Of The Root Logical Table In The Associated Dimension (Doc ID 2379127.1)…

Stupid question, why not make sure that the dynamic list always includes the admin e-mail account(s)? Or better yet, get the e-mail administrator to set you up an account for the purpose and give your admins proxy rights to it so they can look when they need to check.

Hi, Fact1 -> Dimension1 Fact2 -> Dimension2 Then Fact1 -> Dimension2 Will work provided the keys that join the dimension to the facts both have a FK in the fact that joins to the unique key in the dimension, with the LTS levels set appropriately to the appropriate dimension hierarchy level. In my simplified example you…