cancel
Showing results for 
Search instead for 
Did you mean: 
Tristan
Employee
Employee
Status: New

We are running into 3 things that force us to add an extra layer of modeling because we have to calculate/fix data from Fusion before we can push it to Synapse.  These require MV's which add more steps to the pipeline.  If these three items could be addressed, we would save a lot of additional configuration and scheduling time:
 

  1. Support multiple last_update_date columns - the PVO's often have many. This would allow for incremental loads based on changes to multiple PVO sources
  2. 1753 dates conversion - we have a customer that defaults dates to 1753 or other dates (1 AD) in place of nulls.  We need an MV currently to null these out.
  3. 4000 character limit, set to string - Currently have to separate the data into multiple columns with an MV.
2 Comments
RADSr
Captain
Captain

I don't have direct experience w/ Fusion, but can those last two be done by editing the SQL?  

CASE WHEN year(date) = 1753 THEN NULL ELSE year(date) END 

and 

substring(field, 0, 3999) AS Text1, substring(field, 4000, 7999) AS Text2, etc ? 

 

 

Tristan
Employee
Employee

There is not SQL to edit when you pull data from UCM.  UCM holds flat files.