Github user jkbradley commented on the issue:
https://github.com/apache/spark/pull/17090
@MLnick OK I think I misunderstood some of your comments above then. I
see the proposal in SPARK-14409 differs from this PR, so I agree it'd be nice
to resolve it. We can make changes to this PR's schema as long as it happens
soon.
Here are the pros of each as I see them:
1. Nested schema (as in this PR): ```[user, Array((item, rating))]```
* Easy to work with both nested & flattened schema
(```df.select("recommendations.item")```) (AFAIK there's no simple way to zip
and nest the 2 columns when starting with the flattened schema.)
2. Flattened schema (as in SPARK-14409): ```[user, Array(item),
Array(rating)]```
* More efficient to store in Row-based formats like Avro
I'm not sure if there's a performance difference in the formats when stored
in Tungsten Rows. I think not, but that'd be good to know.
---
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
