I have a use case where I start partitioning my data based on a key. All data belonging to a key stays on one partition. e.g.
Key1 -> Partition1 Key2 -> Partition2
Now After sometime I may need to merge the key1 and key2. Hence I just want to transfer all data from key2 to key1 i.e. from partition2 to partition1. Is it possible in apache spark? I don't want to repartition everything as it is costly for me.