For the complete documentation index, see llms.txt. This page is also available as Markdown.

Scala Tips for updateStateByKey

In the previous section on udateStateByKey, following we need to brush up some Scala methods:

Specifically:

val updateFunc = (values: Seq[Int],state: Option[Int])=>{
val currentCount = values.foldLeft(0)(_+_)
val previousCount = state.getOrElse(0)
Some(currentCount + previousCount)
}

A few methods we need to discuss here:

values.foldLeft(0)(_+_)

In fact, there are follow fold* methods for any collections (such as List):

fold(Initial:T)(func)

foldLeft(Initial:T)(func)

foldRight(Initial:T)(func)

Tips: fold methods above are very similar to reduce method, except fold methods allow a initial value as argument, while reduce does not.

state:Option[Int]

this means argument state is an integer, but can be None. In Scala, it is denoted as Option[Int]

If a variable in Scala may be None in value, it must be declared as Option[T] (T is a datatype placeholder, such as Int, String etc)

An Option[T] can be either Some[T] or None object, which represents a missing value.

If a variable needs to be computed with, for example having data type Option[Int], to avoid computing a None value, an Optional variable has a method called getOrElse(default value), such as

state.getOrElse(0)

For the above function updateFunc, let's test it:

Then create a list x, and y:Option[Int], assigned to None, then call updateFunc

Now, replace foldLeft(0)(_+_) with fold(0)(_+_), foldRight(0)(_+_) and reduce(_+_), and test them to reinforce understanding.

Last updated