Skip to content

Relax pandas pin #586

Closed
Closed

Description

@dhirschfeld

Pandas 2.3 has been released. The pin in databricks-sql-python makes it uninstallable alongside the latest pandas:

pandas = [
{ version = ">=1.2.5,<2.3.0", python = ">=3.8,<3.13" },
{ version = ">=2.2.3,<2.3.0", python = ">=3.13" }
]

Activity

  1. jprakash-db commented on Jun 9, 2025

    @jprakash-db
    Contributor

    We will test it once and release the pin. For the time being, you can manually install the required pandas using pip install pandas==2.3.0

  2. stephen-bowser commented on Jul 1, 2025

    @stephen-bowser

    Is there any update on this? Just running pip install pandas==2.3.0 isn't a great option.

  3. dhirschfeld commented on Jul 1, 2025

    @dhirschfeld
    ContributorAuthor

    We will test it once and release the pin. For the time being, you can manually install the required pandas using pip install pandas==2.3.0

    Not if you don't use a package mangeer which blithly ignores constraints.

  4. ohadber commented on Aug 24, 2025

    @ohadber

    Is there any estimation when this ticket will be resolved ?

  5. g3rv4 commented on Aug 25, 2025

    @g3rv4

    Any updates? pandas shipped 2.3.2

  6. cburroughs commented on Sep 30, 2025

    @cburroughs
  7. branchv commented on Oct 27, 2025

    @branchv

    This now prevents python 3.14 users from installing pre-built wheels, forcing a slow source build

  8. g3rv4 commented on Nov 17, 2025

    @g3rv4

    We will test it once and release the pin. For the time being, you can manually install the required pandas using pip install pandas==2.3.0

    @jprakash-db is there any way we can help with this test?

  9. john-semble commented on Dec 2, 2025

    @john-semble

    Seconded @g3rv4. @jprakash-db is there any support the community can give to help this happen?

    Anyone using poetry for package management is unable to use databricks-sql-python with Python 3.14 unless they change their CICD/build environment to be able compile pandas from source.

  10. dhirschfeld commented on Dec 3, 2025

    @dhirschfeld
    ContributorAuthor

    I find it pretty astonishing that a product that pitches itself as a tool for analytics / data-science doesn't support the latest version of the most popular data-science package. Equally astonishing is the fact that depending on this product prevents you from using the latest Python version.

    I will absolutely be using this as an example to argue against continued use/investment in Databricks at my organisation.

  11. samikshya-db commented on Dec 10, 2025

    @samikshya-db
    Contributor

    Hi @dhirschfeld ,
    I completely understand the frustration :/ Apologies for the same. We are working to get this merged and released soon (in #720 ). Once released, you will be able to use Py 3.14, pandas 2.3.x. We have now added CI checks for the new dependencies too.

  12. rad-pat commented on Dec 16, 2025

    @rad-pat

    Unfortunately @samikshya-db #720 only relaxes the pandas pin for Python 3.14. Since we are running 3.12, the change in that PR will not help, this still needs a fix.

  13. samikshya-db commented on Dec 18, 2025

    @samikshya-db
    Contributor

    Hello folks, this is being worked on here : #721

    Rest assured that your concerns will be resolved in the next release. Thank you for your patience.

  14. samikshya-db commented on Dec 19, 2025

    @samikshya-db
    Contributor

    Hello everyone, just FYI : this is released as part of the release yesterday v4.2.3. Let us know if you face issues. Closing this for now.

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

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions