Struct dataflow_types::client::partitioned::PartitionedClientState

pub struct PartitionedClientState {
    uppers: HashMap<GlobalId, MutableAntichain<Timestamp>>,
    peek_responses: HashMap<u32, HashMap<usize, PeekResponse>>,
    parts: usize,
}

Maintained state for sharded dataflow clients.

This helper type unifies the responses of multiple partitioned workers in order to present as a single worker.

Fields

uppers: HashMap<GlobalId, MutableAntichain<Timestamp>>

Upper frontiers for indexes, sources, and sinks.

peek_responses: HashMap<u32, HashMap<usize, PeekResponse>>

Pending responses for a peek; returnable once all are available.

parts: usize

Number of parts the state machine represents.

Implementations

impl PartitionedClientState

pub fn new(parts: usize) -> Self

Instantiates a new client state machine wrapping a number of parts.

pub fn observe_command(&mut self, command: &Command)

Observes commands that move past, and prepares state for responses.

In particular, this method installs and removes upper frontier maintenance.

pub fn absorb_response(
    &mut self,
    shard_id: usize,
    message: Response
) -> Option<Response>

Absorbs a response, and produces response that should be emitted.