Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/11436#issuecomment-190500180
There's one corner-case which this patch doesn't handle correctly: if we
are trying to cache a block, start unrolling the iterator, and then conclude
that the iterator won't fit in memory **and** we can't drop it down to disk,
then we need to have a way to keep the values that we've already computed.
TL;DR: the memoryStore itself now needs a `maybePutAndReturnIterator` method in
order to preserve the old behavior.
---
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]
