Incompatibility of apache-airflow-providers-google with the latests pandas versions (2.2+)

[edserbin]

Hi Airflow community,
The official list of requirements for https://pypi.org/project/apache-airflow-providers-google/#cross-provider-package-dependencies includes the following version constraint for pandas: >=2.1.2,<2.2

Is it allowed to use a higher version of pandas with this library?
We need to use version 2.2+ of pandas because it includes a bug fix that our business logic depends on:
https://github.com/pandas-dev/pandas/pull/54428/files

To install Python dependencies, we use Poetry. Poetry raises an error related to apache-airflow-providers-google when we try to bump the pandas version. Below is the error message from Poetry for the latest (15.1.0) version of apache-airflow-providers-google:

Because apache-airflow-providers-google (15.1.0) depends on pandas (>=2.1.2,<2.2)
and airflow depends on pandas (2.2.0), apache-airflow-providers-google is forbidden.
So, because airflow depends on apache-airflow-providers-google (15.1.0), version solving failed.

[potiuk]

The upcoming version of the provider already has the limitation removed https://github.com/apache/airflow/blob/main/providers/google/pyproject.toml#L131
-> should be released within next few weeks - look at devlist and possibly even help testing rc candidate when it's out

Also you can always try to force installation with breaking dependencies (not sure how to do it in poetry. It has a chance to work - but you have to test it yourself.

[edserbin]

Thanks for your answer!
I hope a new version without the pandas restriction will be released soon.