Uploaded image for project: 'Apache Arrow'
  1. Apache Arrow
  2. ARROW-76

Revise format document to include null count, defer non-nullable arrays to the domain of metadata

    Details

    • Type: New Feature
    • Status: Resolved
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 0.1.0
    • Component/s: Format
    • Labels:
      None

      Description

      After many mailing list discussions and a first draft of metadata serialization (https://github.com/apache/arrow/blob/master/format/Message.fbs), I propose we strike the notion of required fields from the physical memory layout and instead make this a property of the metadata (see the nullable member of Field in the serialized schema metadata).

      Note that Arrow implementations will be free to create specialized containers for data having nullable = false in their metadata. It will be the implementations' responsibility to validate that null_count == 0 at IPC boundaries when interacting with "untrusted" data.

        Attachments

          Activity

            People

            • Assignee:
              wesmckinn Wes McKinney
              Reporter:
              wesmckinn Wes McKinney
            • Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: