-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/41505/#review111769
-----------------------------------------------------------


I had an offline discussion with Peeyush Bishnoi over this. Here is the summary 
of the discussion.

1. We can use vertex id as the unique parameter
2. Although there is no direct API leveraging timestamp, there is a REXSTER api 
exposed using which anyone can query any property, so all properties are in a 
way public contract.
3. Using same property for nominal time can lead to issues if users don't 
migrate incorrect data. e.g. if users query over a time range which has both 
incorrect data and correct data, users will get several incorrect results. 
Using a second property will return no such wrong results.
4. If we decide to use the same property then it's fine but it will be helpful 
to call out the change explicitly so that users are aware. 

Peeyush wants to try some more approaches, will get back on this after that.

- Ajay Yadava


On Dec. 18, 2015, 6:48 a.m., Peeyush Bishnoi wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/41505/
> -----------------------------------------------------------
> 
> (Updated Dec. 18, 2015, 6:48 a.m.)
> 
> 
> Review request for Falcon.
> 
> 
> Bugs: FALCON-1661
>     https://issues.apache.org/jira/browse/FALCON-1661
> 
> 
> Repository: falcon-git
> 
> 
> Description
> -------
> 
> Similar timestamp is getting added for Falcon instance vertex in GraphDB
> 
> 
> Diffs
> -----
> 
>   
> common/src/main/java/org/apache/falcon/metadata/InstanceRelationshipGraphBuilder.java
>  b709857 
>   
> common/src/main/java/org/apache/falcon/workflow/WorkflowExecutionContext.java 
> f206ff1 
>   
> common/src/test/java/org/apache/falcon/metadata/MetadataMappingServiceTest.java
>  29f933d 
> 
> Diff: https://reviews.apache.org/r/41505/diff/
> 
> 
> Testing
> -------
> 
> yes
> 
> 
> Thanks,
> 
> Peeyush Bishnoi
> 
>

Reply via email to