I have to verify that olap cube data and the data from relational tables from where a cube is built is correct.
And I will do so by writing the TSQL queries and compare the values with that of cube.
But, I got stuck in the course of determining which columns are used for measure. How do I figure out which columns are used for measures?
Help appreciated!
You need to look at the cube metadata.
For SSAS2005, take a look at the DSV (data source view) and mappings to dim and fact table values behind the scenes. This should allow you to see what is going on. If you don't have a project you can reverse engineer it using the 'import anslysis services template' (or some such) option from the new project dialog in BIDS.
Calcuated measures are defined in the cube script. If you have a reverse-engineered cube or cube project you can open the cube and see this in the 'calculations' tab.
For AS2000 you can open the cubes on the server (assuming sufficient permissions) and look at the mappings there. There is a tool called OLAPScribe that will help you do this for AS2000. Alternatively you can run a trace on the source database and capture the SQL generated by the cube as it is processed.
Related
Is there any command to refresh cube schema (Structure) every day before processing the SSAS tabular cube?
Requirement: Few column names in an object keeps changing(renaming) so, if we rename those columns in DB view, it should automatically reflect in Cube after overnight processing.
I tried adding ‘select * from an Object’ in cube table properties and processed cube.
Later, I tried to rename column name in view then processed cube, it failed due to different column name.
Is it possible to dynamically refresh schema without manually changing structure via solution and re-deploy?
Please suggest
This answer is not going to solve your issue, however best practice would be NOT to handle these changes in the SSAS Data Model. This should be a consistent layer for your users. I would handle these changes in the data base or source system end, maybe checking if a column exists, and if not adding a dummy row or some other process.
I have access to db2.
I also have access to some Cognos Cubes via TM1 explorer.
Is it possible to find out the source tables for the cubes ?
For example, can I see what tables are used for the FST cube shown below ? and if possible even how the tables are joined etc
. ?
To answer your question, you need access to the TurboIntegrator process that load your cube.
In your example, since it's an external source (Db2), the cube is loaded with a TurboIntegrator Process (process object in green). It's seems that you don't have the security access to view the process so you're not able to see the data source.
Hope it helps !
First Post! am kinda feeling the pressure :)
I've created a multiple option filter for my dashboard using a set (with include all members). which is great for my sheets which use that datasource as primary, but when it's the secondary source I've hit a brick wall - I cannot see (or find any reference in searches) how to use the filter. Any calculated dimensions I've seen reference In/Out. Is there a way round this or something I'm missing?
thanks.
You can't do it when it's based on the secondary source in a data-blending scenario. Currently Tableau only allows for it to be applied to the primary. We are hoping this changes in version 10.0.
I am having an issue finding links within Crystal Reporting. I am trying to link 2 tables. However this seems to be a long task as I battle to find any links.
Is there a method that should be followed, as the most that have the same names to not link the tables correctly.
You should take a look on AX Data Dictionary Tables for the involved tables, then check the Relations node of the table. Any direct relation is usually on the child table. This of cause requires developer rights to the target AX system.
Remember to join on DataAreaId and PartitionId as well. Always check if any appropriate index exist.
Also the Reverse Engineering Tool may be of help.
I would like to know if it's possible to "unprocess" a cube in SSAS. I want to do this so I can perform a "synchronization" between two differents servers. This (I think), will remove all the data and left the metadata to be synchronized? Am I right?
Do you know a way to do this?
Thanks!
There is the ability to unprocess, although that'd leave your cube unprocessed. Right-click database, select process, in the next window change process type to Unprocess and "Script to new window" in top-right.
If you just want a schema though, can you not script the "Script Database as Create" XMLA? Then drop your target and run the Create XMLA against the target server?