KB313409: Cross joins appear in SQL statement when running a Visual Insight dashboard with a live connection and adding a filter object to the filter visualization in MicroStrategy Web 10.3

SYMPTOM: 
Cross joins appear in the SQL statements when running a Visual Insight dashboard with live connection and adding a filter object to the filter visualization in Strategy Web 10.3.
STEPS TO REPRODUCE:

1. Enable the SQL trace log, by following KB13189: How to trace the SQL generated by MicroStrategy Engine 9.x .
2. In Strategy Web, import data of 4 tables from a data source with live connection.
3. Save the imported data cube. 
4. Create a Visual Insight dashbaord based on the cube.
5. Add some attributes to filter panel with a checkbox style.
6. Add some other attributes to filter visualization with checkbox style. Make sure the attributes on the filter visualization and filter panels are from different tables and unrelated.
7. Go to SQL trace log, note that there are cross joins in the SQL statement of the dashboard, like shown below:

8. 
   select	distinct a13.CAL_YEAR  CAL_YEAR,
   	a11.SUBCATEGORY_SEQ_NBR  SUBCATEGORY_SEQ_NBR,
   	a11.SUBCATEGORY_DESC  SUBCATEGORY_DESC,
   	a12.REGION_NAME  REGION_NAME,
   	a11.CATEGORY_SEQ_NBR  CATEGORY_SEQ_NBR,
   	a11.CATEGORY_DESC  CATEGORY_DESC
   from	dbo.TMP_SUBCATG_DIM	a11
   	cross join	dbo.TMP_CUST_DIM	a12
   	cross join	dbo.TMP_TIME_DIM	a13
   where	(a13.CAL_YEAR in (2016) and a11.CATEGORY_SEQ_NBR in (1, 3))
   

   
   

 
CAUSE: 
This is a known issue with Strategy Web 10.3.
 
ACTION: 
This issue has been addressed in Strategy 10.6 a "Feature Release". Before upgrading to the previous version please review the differences between a "Feature Release" and a "Platform Release". These differences are detailed here. For any questions or concerns regarding the "Feature Release" and the "Platform Release" please contact Strategy Technical Support. 
 
WORKAROUND: 
Either by moving all the filters to the filter panel to use filter panel alone or, moving all the filters to filter visualization to use filter visualization alone, this will avoid cross join to happen.