Github user jkbradley commented on the issue:

    https://github.com/apache/spark/pull/15770
  
    Yep, that's correct.  Everyone, please let me know if you disagree.
    
    Also, if we do go with Option 2 above, then the input schema could be a few 
possible things:
    * list of (neighbor ID, weight)
      * pro: ensures # neighbors = # weights
      * con: nesting can be annoying to work with
    * (list of neighbor IDs, list of weights)
      * pro: Easier to support unweighted graphs by omitting list of weights.  
Avoid nesting.
      * con: could have invalid input with mismatched numbers of neighbors, 
weights
    * Vector (of length # neighbors)
      * pro: Familiar data type.  Ensures # neighbors = # weights.
      * con: If we later support passing raw features which are used to compute 
a similarity matrix within PIC, then this could confuse users.
    
    I'm ambivalent about these schema options.  What do you think?


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org

Reply via email to