Skip to content

Float16 treats Nan, +0 and -0 differently than Double. #3406

Description

@HuaHuaY

Describe the bug, including details regarding any error messages, version, and platform.

It should not create column index when the min or max value is NaN. And it should adjust the min or max value when meets +0 or -0.

Component(s)

No response

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions