Thank you for your response @bednar!

Well, looking at other use cases within alpakka connectors, I would like to see something along these lines:

A plain sink, which may or may not do under the hood batching:
Sink[Point[T], Future[WriteResult]] where WriteResult would be an indicator of an operation, perhaps Done is a good starting candidate if no other information is to be transmitted
An explicit batched sink:
Sink[Seq[Point[T]], Future[WriteResult]]
A simple flow:
Flow[Point[T], WriteResult, Mat] where this WriteResult could be different from the one in sink, here perhaps carrying more information, such as input + response
Batched flow
Flow[Seq[Point[T]], WriteResult, Mat] with "atomicity" guarantees with a batch, meaning if the input (Seq[Point[T]]) has passed through the flow, then all the elements have been written (or submitted to writing, depending on which delivery guarantees are used)
Flow with context
FlowWithContext[Point[T], C, WriteResult, C, Mat]
Batched flow with context
FlowWithContext[Seq[Point[T]], C, WriteResults, C, Mat]

Edit:
Looking at alpakka-kafka, a unifying type such as Envelope a = Single a | Multi [a] | PassThrough (and their corresponding responses) might be more ergonomic, and then simple and batched flow(withContext) and could be replaced with a "flexiFlow(withContext)": Flow[Envelope[T, PT], Results[T, PT], Mat]

Write API for Scala seems to be missing #125

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions