Semi-Additive measures
Hi - i have a data source on premise that is OLAP. I know that using Pulse I cannot access OLAP sources and I must access relational data sources only (in our case MS SQL Server) - i have done this and I am faced with an issue.
Currently, the OLAP cube is queried in Excel and embedded in the OLAP there are some measures which are defined as "Last Non-Null". This allows us to roll up stuff like "outstanding customer balance" for each week to the month without totalling all weeks (which is incorrect).
When we access relational databases, is it possible to add an option to roll up against time measures like the "Last Non-Null" or - allow Pulse to access an Analysis Services cube?
1 reply
-
The relational querying engine does not support these types of aggregations (SQL itself has no such logic). We may add it in the future.
Separately, accessing Analysis Services on premise, via Pulse, is heavily complicated due to the issue of authentication via Kerberos (which MS OLAP requires). At the moment we have avoided this capability, because it would require a very messy integration of the security running your cloud instances with on-premise to work fully.
If you are desperate to connect, we would suggest setting up a VPN between the cloud and your on premise network; adding the cloud machines to the AD; and then using AD authentication in Pyramid to 'naturally' flow Kerberos authenticated requests to your on-premise cubes for querying. A big story indeed!