Github user ggevay commented on the issue:
https://github.com/apache/flink/pull/2094
> As for this pull request I suggest to update the false Javadoc from the
streaming API and document in the concepts that the aggregation field accessors
and the keyselectors are slightly different due to there semantics.
I've fixed the javadoc now in 154352ba879350bb1ab10b799456a67eb5819375.
As for the differences between the key selectors and aggregations (which
use `FieldAccessor`): I tried to consolidate the supported field expressions
between the two in 7c2dceeb87b06963cc2618312f5f1039387003ad. Now the
aggregations should accept any field expression that selects a single field and
would be a valid key selector, and some more (selecting array elements).
As for documenting that the aggregations don't work on a field that is
still some composite type, I'm not sure if this is necessary, since this should
be pretty obvious to the user.
> The batch API aggregations do not have the overloaded version where you
could pass a field accessor string.
Here is the Jira for this: https://issues.apache.org/jira/browse/FLINK-4575
---
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.
---
