Skip to content

TempSource: dumping data into a location to persist between Jobs #390

@johnynek

Description

@johnynek

In page rank, we run a job, write out a value, and before we start the next Job/Flow, we read the value and see if we should stop.

Currently, our implementation is ugly and requires the user to control that. What would be better is something like:

  val error = Temp() // can only write ONCE in the set of flows.
 // some place later in the same "family" of jobs:
  val result = error.read

So we need to keep a map of all the open temp files, which jobs write into them (we want write-once semantics) and then clean them up just before the Tool exits.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions