Hi Wei,
what do you mean be "windowedStream"?

The result of
dataStream.join(otherStream).where(<key selector>).equalTo(<key selector>)
expects a window to be specified. In each window, based on the time and window 
characteristics you defined, both sources will collect elements that fit into 
the window and, at its end, will complete the join operation by calling your 
user function "for each combination of elements with the same key in a window" 
(see JoinedStreams.java). The result is a (non-keyed) data stream again.

Of course, you can always apply additional window operations afterwards if 
that makes sense.


Nico

On Tuesday, 25 July 2017 09:15:31 CEST xie wei wrote:
> Hello Flink,
> 
> 
> 
> one question about join operator:
> 
> Flink Docu said "Join two data streams on a given key and a common
> window",=  does "common window" mean that the joined stream is
> windowedStream?
> 
> If it is normal datastream (not windowed), could the stream be windowed
> aga= in with different window definition as the window specified by join?
> 
> 
> 
> 
> 
> Thank you and best regards
> 
> Wei

Attachment: signature.asc
Description: This is a digitally signed message part.

Reply via email to