Github user tammymendt commented on the pull request:
https://github.com/apache/flink/pull/605#issuecomment-109990676
Hey!
So I solved the NullPointer and as far as I can see it is not bothering any
more. Given the non-deterministic nature of the algorithms, some tests fail
sometimes. I am looking into this. I am considering leaving out tests for the
accuracy of the algorithms (in the Accumulator test), given that they are
already programmed in other test classes. This should avoid these "algorithm
accuracy" related errors in the OperatorStatisticsAccumulatorTest class.
Also, I have been refactoring this code to avoid using the conditionals
that check which statistics are being collected in the "process" and "merge"
methods. This is based on a remark that @fhueske made way earlier in the PR.
Should I try and make a PR for the refactored code, or rather leave this in its
current state, and wait until it is merged until I make a PR for the refactored
version?
Cheers!
---
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.
---
