[GitHub] metron issue #863: METRON-1347: Indexing Topology should fail tuples without...

2018-04-13 Thread mmiklavc
Github user mmiklavc commented on the issue:

https://github.com/apache/metron/pull/863
  
@ottobackwards created this to track it - 
https://issues.apache.org/jira/browse/METRON-1524


---


[GitHub] metron issue #863: METRON-1347: Indexing Topology should fail tuples without...

2018-04-12 Thread ottobackwards
Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/863
  
I think this is a mess that we will need to sort out, but I won't hold this 
up for that.
I'm +0 I guess.  


---


[GitHub] metron issue #863: METRON-1347: Indexing Topology should fail tuples without...

2018-04-12 Thread mmiklavc
Github user mmiklavc commented on the issue:

https://github.com/apache/metron/pull/863
  
It seems this PR is meant to address the global case, since the writer is 
general purpose. Per @cestella's comments, I don't know that we can include 
original_string. I do think there's value in minimally ensuring the timestamp 
and source.type fields are validated in the near term. It seems that addressing 
a broader validation strategy might be in order, but maybe we could handle that 
separately as a compromise? Thoughts @ottobackwards?


---


[GitHub] metron issue #863: METRON-1347: Indexing Topology should fail tuples without...

2018-03-19 Thread ottobackwards
Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/863
  
What I think would be better would be to have a readme about the required 
fields and refer to that document from these other documents


---


[GitHub] metron issue #863: METRON-1347: Indexing Topology should fail tuples without...

2018-03-19 Thread ottobackwards
Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/863
  
The parser topology PRODUCES the original_string.  By that very action, and 
it's requirement, it can only be used and required by downstream usages either 
from rest ( post indexing ) or during enrichment.  It seems inconsistent to 
validate without including that, if what you want to do is validate the entire 
set of required fields.  If you only want to handle what is required at a given 
step, but not what is required 'overall' then this is OK, but something doesn't 
seem right here conceptually.


---


[GitHub] metron issue #863: METRON-1347: Indexing Topology should fail tuples without...

2018-03-18 Thread justinleet
Github user justinleet commented on the issue:

https://github.com/apache/metron/pull/863
  
@mmiklavc @cestella @ottobackwards  Is this good to be merged in?


---