Skip to content

Reduce the number of copies in binary operations in DataFrame #5663

Closed as not planned

Description

@pgovind

We make copies of columns in the binary operation APIs mostly to reuse existing code. PrimitiveDataFrameColumnArithmetic should be improved to work on columns of different types. Once that is done, we likely won't need any more copies.

Activity

  1. transferred this issue fromdotnet/corefxlabon Mar 6, 2021
  2. rosebyte commented on Oct 13, 2026

    @rosebyte
    Member

    Triage: we don't want to invest in performance of this library at the moment.

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

    area-DataFrameMicrosoft.Data.Analysis DataFrame APIsyellowmaintenance only

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions