In addition to the sendMsg
and mergeMsg
methods, aggregateMessages
can also take an optional argument TripletFields
, which indicates what data is accessed in EdgeContext
. The main reason for explicitly specifying such information is to help optimize the performance of the aggregateMessages
operation.
In fact, TripletFields
represents a subset of the fields of _EdgeTriplet_
and it enables GraphX to populate only those fields that are necessary.
The default value is TripletFields.All
, which means that the sendMsg
function may access any of the fields in the EdgeContext
class. Otherwise, the TripletFields
argument is used to tell GraphX that only part of EdgeContext
will be required so that an efficient join strategy can be used. All possible options for the TripletFields
are listed as follows:
TripletFields.All
: This option exposes all the fields (source, edge, and destination)TripletFields.Dst
: This one exposes the destination and edge fields but not the source field...