Re: Join with Default-Value

Posted by Gábor Gévay on Feb 10, 2017; 12:05pm
URL: http://deprecated-apache-flink-user-mailing-list-archive.369.s1.nabble.com/Join-with-Default-Value-tp11566p11567.html

Hello Sebastian,

You can use DataSet.leftOuterJoin for this.

Best,
Gábor




2017-02-10 12:58 GMT+01:00 Sebastian Neef <[hidden email]>:

> Hi,
>
> is it possible to assign a "default" value to elements that didn't match?
>
> For example I have the following two datasets:
>
> |DataSetA | DataSetB|
> ---------------------
> |id=1     | id=1
> |id=2     | id=3
> |id=5     | id=4
> |id=6     | id=6
>
> When doing a join with:
>
> A.join(B).where( KeySelector(A.id))
>         .equalTo(KeySelector(B.id))
>
> The resulting dataset is:
>
> |(DataSetA | DataSetB)|
> ---------------------
> |(id=1    | id=1)
> |(id=6    | id=6)
>
> What is the best way to assign a default value to the elements id=2/id=5
> from DataSet A. E.g. I need a result which looks similar to this:
>
> |(DataSetA | DataSetB)|
> ---------------------
> |(id=1    | id=1)
> |(id=2    | Default)
> |(id=5    | Default)
> |(id=6    | id=6)
>
> My idea would be to get the missing Elements from DataSetA by .filter
> with (DataSetA|DataSetB) and then do a .union after creating a tuple
> with a default value. But that sounds a bit over-complicated.
>
> Best regards,
> Sebastian