Page History

...

single stream					multiple streams
data transformation				other	data transformation
tuple-by-tuple (KStream -> KStream)		aggregation			tuple-by-tuple (i.e., joins)
same key	new key	non-windowed (KStream -> KTable)	windowed (KTable -> KTable<<W,K> V>		non-windowed KStream-KTable	windowed KStream-KStream
filter (1:[0,1])		aggregateByKey	aggregateByKey	print	join (as of 0.10.2)	join
filterNot (1:[0,1])		reduceByKey	reduceByKey	writeAsText	leftJoin	leftJoin
	selectKey (1:1)	countByKey	countByKey	foreach		outerJoin
mapValues (1:1)	map (1:1)			through
flatMapValues (1:n)	flatMap (1:n)			to
transformValues (1:1 + state)	transform (1:1 n + state)			branch
	process (1:n 0 + state)

KTable API

KTable currently offers the following methods which do have different implication with regard to (internally) created topics and RocksDB usage.

Jira

...