invalid column name

  • 16 March 2023
  • 5 replies


Hello everyone,

After the odx transfer task for API, when I click on the preview, I get the “invalid column name” error. This warning also occurs while executing in the data warehouse and stops execute.  What should I do? I wanted to use aliases but I have 575 columns.

I tried to re-deploy the table with differential deployment disabled. 


Version: 6143.1


My column name example is below: it is very long




Best Regards


Best answer by ebay 31 March 2023, 14:50

View original

5 replies


--solved-- how?

I deleted the '_' in the path separator and it worked. But column names didn’t change. Why? Can someone explain to me?

Userlevel 3
Badge +3

Which provider are you using ?  For root cause, can you configure logging for the provider data source in the portal (example image below) and generate two log files?  One for the non-working scenario (original column name) and then for the working scenario (updated column name)




Provider type:

Managed ADO.NET

Provider version:


Provider name:


Userlevel 6
Badge +5

Please see the following documentation regarding the path separator property for the JSON provider

It appears that the path separator controls the table name based on the file name. I am not sure how changing this property would resolve an “invalid column name” error. Also, the default value for the path separator is ‘_’, so deleting '_' in the path separator property should in theory not have made a difference, since the default would be used which is the same. Perhaps you synchronized the data source after changing the path separator, which actually solved the “invalid column name” error (rather than changing the path separator property).



it was a JSON file and it merged into a single column by changing the schema. Finally, I got a column by subject, not by dates. (Exp instead of ‘WS10M_RANGE_202109’  I got WS10M_RANGE)

Best Regards