not compatible with pandas >= 2.0.3

[randomrandomguy12345]

I had hard time finding the list of pandas versions that are compatible with the package.

Anyway, the package isn't compatible with pandas 2. This is a problem, as pandas 2 did a few changes here and there, and this breaks the code. Downgrading pandas creates a lot of trouble with big projects. Is there a prospect that in near future the package will be compatible with the newer versions of pandas?

[rsheftel]

Making it pandas 2.0 compatible is the next item on the task list. For now the latest PyPi release has pandas < 2.0 as a requirement

[randomrandomguy12345]

Thanks!

[LongSimple]

Thank you

[programmingwithalex]

@rsheftel is there any update on when the update will be deployed? I'll attempt to fork the repo and make a pull request if there's been no progress yet.

[nathanramoscfa]

I forked the project, made this change to allow pandas greater than 2.0:

https://github.com/nathanramoscfa/pandas_market_calendars/commit/0c15491795280123a0b965fb9f0a8b9d65511075

I saw that there were some commits to this project to support pandas 2.1.x so I am trusting that when I install this forked version of the project, it'll work with pandas>=2.0. Well, I installed this forked version to my environment, and reran my Python application, and it worked with no problems. For reference, the pandas compatibility issue came up for me when I tried to build the docs for my application on ReadTheDocs, which would fail, because of this issue.

[krazykoder]

@nathanramoscfa - I am a noob. Appreciate any input. Can you write a few lines of instruction on how to build the wheel and install with pip after I git clone your fork. I am having the same issue - some of my code required pandas >2.0.

[nathanramoscfa]

@nathanramoscfa - I am a noob. Appreciate any input.

Can you write a few lines of instruction on how to build the wheel and install with pip after I git clone your fork. I am having the same issue - some of my code required pandas >2.0.

Starting in command prompt, activate your Python environment. I like to use anaconda, so for me it's like this (replace environment name with the actual name of your environment without the <>):

conda activate <environment name>

Navigate to your folder where you store all your Python projects, could be something like this:

cd C:/Python Projects/

Clone my forked version there:

git clone https://github.com/nathanramoscfa/pandas_market_calendars.git

Then navigate to the project's root directory:

cd pandas_market_calendars

Then install as a package to your environment.

pip install .

It should then install the cloned forked project to your Python environment and now when you import pandas_market_calendars into your application, it'll be using my forked version.

Whenever the pandas_market_calendars project updates the PyPi package with the changes we're discussing here, you would then use this command to upgrade the pandas_market_calendars back to the official pip installed version.

pip install --upgrade pandas_market_calendars

I just simply needed a solution now, couldn't wait for the devs to update the official PyPi package. But when they do I'll revert back to that version instead of my forked version.

[rundef]

FWIW I'm using a slightly older version (4.3.1) with pandas 2.1 and haven't encountered any issue

[rsheftel]

Thanks. This is going to require some investigation as this library uses a lot of the deep internals of pandas to work and much of the pandas holiday functionality is undocumented.

[kentonself]

Hi.

I am running pandas_market_calendars 4.3.3 with pandas 2.2.0. Everything runs fine, except, I get the following message when I call

get_calendar("NYSE").open_at_Time(schedule=x, timestamp=y)`

FutureWarning: Downcasting behavior in replace is deprecated and will be removed in a future version. To retain the old behavior, explicitly call result.infer_objects(copy=False). To opt-in to the future behavior, set pd.set_option('future.no_silent_downcasting', True)

I am assuming this is related to this issue/project.

[rsheftel]

v.4.4.0 is published to PyPi and works with Python 3.12 and Pandas 2.2.0