I am using PowerBI desktop and connecting to a SSAS tabular model cube. This is working just fine, except there are three measures missing from the list of fields.
Through experimentation, I was able to determine that any measure with a MEDIAN or MEDIANX function is not being brought into PowerBI. If I use SUM, the measure will appear. I made sure to check for hidden measures but anything with those MEDIAN functions are nowhere to be found.
These are simple measures, similar to:
Median of X:=MEDIAN([X])
It would appear PowerBI is filtering out Medians on purpose, but I can't figure out why. I suppose I could make my own Median measure in the PowerBI desktop, but my clients want to be able to easily grab the measure from the cube... which kind of makes sense because that's why we built the cube in the first place, right?
Any ideas on how to fix this? Any help will be greatly appreciated.
UPDATE:
I have tried adding the measure three different ways:
Median Measure1:=MEDIAN([Column])
Median Measure2:=MEDIAN(Table[Column])
Median MeasureX:=MEDIANX(Table, [Column])
All three appear in the measures when I load the data source to a PivotTable. They all work identical.
I also connected to this data source in SSRS Report Builder and I am able to see all three measures.
I then connect to the datasource live in PowerBI Desktop. The measures are nowhere to be found. I can search for "median" and I receive no results. If I view hidden fields, they are still nowhere to be found.
I am using the following PowerBI Desktop version:
2.50.4859.502 64-bit (September 2017)
I will also add that I have other aggregate measures using the same table/column that are appearing fine in Power BI.
Our SSAS Tabular models are using SQL Server 2012 RTM (1100) compatibility level. Would this affect the measure in PowerBI?
This question was posted to the PowerBI forums and I will update this question if I get an answer on there.
Related
We have VS 2019 for SSAS development (with installed SSDT).
I would like to create aggregations for measures in order to increase query performance (cube with tabular model).
Could you tell, please, where the window "Aggregation" is located in VS?
P.S. on the screen you can see a snapshot from the video on youtube where i highlighted the ribbon (and button) i am searching for.
From the screenshot, it looks like, You are looking into a video on SSAS (multi dimensional). It is having aggregations. For SSAS (Tabular), there are no custom aggregations.
You can see the difference between SSAS (multidimensional) Vs SSAS (Tabular) here
Feature
Multidimensional
Tabular
Aggregations
Yes
No
I have a MDM cube with fact having just 84K records. Simple STAR schema structure. All regular relationships and no calculations or KPIs.
While trying to browse the cube by pulling 15-20 attributes from various dimensions with no filters applied, cube gives result in less than 10-12 seconds, which is awesome.
While reading the cube through Excel pivot, as soon as the attribute count goes beyond 8 or so, cube becomes slow to a point wherein if we bring in even a single attribute in column labels, it takes forever to respond, sometimes even crashes.
Has anyone faced such issue before and what could be a solution or workaround?
PS: I know for a fact that Excel creates a very complex MDX query to retrieve data from MDM cube which otherwise can be extracted using a simpler MDX.
I've working on SSAS cube creation first time on real data(Although had little hands on experience on AdventureWorksDW08).I also have gone through the details for creating basics cube and all but have little question as to which are the pre-requisite for creating the SSAS cube .e.g (SQL TABLE SHOULD NOT HAVE NULL VALUES) etc.
I'm Tried so many times but it seems that data is not get refresh for certain result columns.
I'm creating what's supposed to be a very simple cube, with just one measure , and about 7 dimensions. Each dimension comes from a different table.
It looks like the standard setup is to always have a folder visible for each separate dimension, and also a separate folder for measures. Normally I can see where that would be desirable, if you have many attributes in one dimension. However, in this case it's unnecessary and adds complexity.
Is it possible to NOT have separate folders for each dimension?
thanks,
Sylvia
Unfortunately it is not possible to avoid folders of dimensions but you can combine multiple measures from different measure groups in a new folder and make original ones hidden. That way you can achieve partially but only for a limited case.
It really depends on what your client is for the cubes.
You see , if you look at things through mgmt studio, or through bids or through excel.. .each will display the dims and measure groups \ measures in their own way. Then there are other tools like old proclarity and performance point or tableau or reporting services which allow more flexibility .
Im guessing you're not using Excel as your front end , as it doesnt have folders as such. Althought the various hierarchies within a dim are grouped together under the dimension. And each measure group lists its own measures grouped together.
Look at the excel layout and see.
aamer
I am using olap4j to query my rolap cube (underlying implementation is pentaho).
I could not found any way to add on-the-fly calculated measurement to the query while using the org.olap4j.query package.(there is a way when using the low level api in org.olap4j.mdx package)
Is there any support of olap4j for calculated measurements that I am missing?
Yosi
So far, calculated members are not a concept supported by the query model. There are other query models available out there; some might support them.
If you are interested in working in that area, or just getting more information, I suggest you join the olap4j mailing list.