I changed it to Tuple2 and that problem is solved.
Any thoughts on this message
*Unapplied methods are only converted to functions when a function type is
expected.*
*You can make this conversion explicit by writing `updateAcrossEvents _` or
`updateAcrossEvents(_,_,_,_,_)` instead of
maybe you can combine the fields you want to use into one field
Something Something 于2020年3月3日周二 上午6:37写道:
> I am writing a Stateful Streaming application in which I am using
> mapGroupsWithState to create aggregates for Groups but I need to create
> *Groups
> based on more than one column in
I am writing a Stateful Streaming application in which I am using
mapGroupsWithState to create aggregates for Groups but I need to create *Groups
based on more than one column in the Input Row*. All the examples in the
'Spark: The Definitive Guide' use only one column such as 'User' or
'Device'. I