Async I/O: preserve stream order for requests on key level

[Juho Autio <ju...@rovio.com>]

I need to make some slower external requests in parallel, so Async I/O
helps greatly with that. However, I also need to make the requests in a
certain order per key. Is that possible with Async I/O?

The documentation[1] talks about preserving the stream order of
results, but it doesn't discuss the order of the async requests. I tried to
use AsyncDataStream.orderedWait, but the order of async requests seems to
be random – the order of calls gets shuffled even if I
use AsyncDataStream.orderedWait.

If that is by design, would there be any suggestion how to work around
that? I was thinking of collecting all events of the same key into a
List, so that the async operator gets a list instead of individual events.
There are of course some downsides with using a List, so I would rather
have something better.

In a nutshell my code is:

AsyncDataStream.orderedWait(stream.keyBy(key), asyncFunction)

The asyncFunction extends RichAsyncFunction.

Thanks!

[1]
https://nightlies.apache.org/flink/flink-docs-master/docs/dev/datastream/operators/asyncio/#order-of-results