Github user yanboliang commented on the issue:
https://github.com/apache/spark/pull/18902
@zhengruifeng DataFrame-based operation is 2~3x slower than RDD-based
operation is a known issue, because of the deserialization cost. If we switch
to RDD-based method, we need to implement our own aggregator to calculate
_mean_ and _median_, this need much more code than calling DataFrame API. BTW,
DF using more compact structure that can reduce memory footprint.
From my perspective, I'd suggest to keep the current DF-based solution. As
it will 5~10 faster than the original implementation. @hhbyyh @MLnick What do
you think about it? Thanks.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
