Github user sachingoel0101 commented on the pull request:
https://github.com/apache/flink/pull/949#issuecomment-134070592
Hey @ChengXiangLi, I have another concern, regarding the seed for sampling.
It doesn't seem to serve its purpose. I tried sampling with fraction three
times with the same seed, however, every time I get different results.
I've been stuck on this problem myself for quite some time. The fix is,
instead of `seed + index`, you should just use `seed`. But then, the sample is
not truly random. Far as I could figure out, splits of data don't arrive at the
exactly same index subtask every time.
---
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.
---
