docs: update recommended pyathena version #24200
Conversation
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
| @@ -22,8 +22,8 @@ A list of some of the recommended packages. | |||
|
|
|||
| | Database | PyPI package | Connection String | | |||
| | --------------------------------------------------------- | ---------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- | | |||
| | [Amazon Athena](/docs/databases/athena) | `pip install "PyAthenaJDBC>1.0.9` , `pip install "PyAthena>1.2.0` | `awsathena+rest://{aws_access_key_id}:{aws_secret_access_key}@athena.{region_name}.amazonaws.com/{ ` | | |||
| | [Amazon DynamoDB](/docs/databases/dynamodb) | `pip install "PyDynamoDB>=0.4.2` | `dynamodb://{access_key_id}:{secret_access_key}@dynamodb.{region_name}.amazonaws.com?connector=superset` | | |||
| | [Amazon Athena](/docs/databases/athena) | `pip install "PyAthenaJDBC>1.0.9"` , `pip install "pyathena[pandas]>=2,<3"` | `awsathena+rest://{aws_access_key_id}:{aws_secret_access_key}@athena.{region_name}.amazonaws.com/{ ` | | |||
There was a problem hiding this comment.
I believe it would be better if we treat this table as an example and point the reader to the files where the most up to date versions can be found such as setup.py and requirements/*. That way, this documentation will be more resilient to version upgrades.
There was a problem hiding this comment.
Agree, should I extend the scope of the current PR, or open a new PR for this?
There was a problem hiding this comment.
Thank you for opening a new PR for it.
There was a problem hiding this comment.
@michael-s-molina
I tried looking into this issue once before. I couldn't figure out why this table had been changed from referencing the optional extras requirements that are managed in the setup.py to the state the table now is in where it no longer specifies the optional extras requirements but instead directly lists the actual driver package names in the documentation table itself.
I think there is a bigger question of how to manage the driver compatibility and support in general. I.e. should the be such a concept of "officially supported and tested drivers" that should have an expectation of always working, vs community-supported and tested drivers?
There was a problem hiding this comment.
That's an interesting question @cwegener. My suggestion would be to create a GitHub discussion and tag some folks such as @dpgaspar @villebro @john-bodley @rusackas @betodealmeida and myself or bring this to the Town Hall.
SUMMARY
The recommended version of pyathena in the doc is
PyAthena>1.2.0, which is not in sync with the version specified in setup.py, which ispyathena[pandas]>=2,<3.If pyathena v3 is installed, it will result in a runtime error
Name "Optional" is not defined. Revert back to v2 resolved the issue and the connection works correctly.Also, other packages lack version constrain in the same doc, should I also update them? Let me know.