Hi everyone, what would be the reason

kedro-datasets>=2

pinned

s3fs

to

["s3fs>=2021.4, <2024.1",]

? I had a great time last night upgrading from 1.x to 3.0 😄 - some

s3fs

versions prior to

2024.1

have some weird stuff going on: https://github.com/fsspec/s3fs/issues/504 and upgrading to

2024.3.1

solved those issues for me. However I had to take that dependency out of my

requirements.txt

to make it work in the end. Also, are there any examples in the docs about the right

pip

syntax to install the different

kedro-dataset

optional datasets? Thanks!

> Hi everyone, what would be the reason

kedro-datasets>=2

pinned

s3fs

to

["s3fs>=2021.4, <2024.1",]

? I'm a big fan of not having arbitrary upper bounds for this reason—https://iscinumpy.dev/post/bound-version-constraints/ is my go-to article on the topic—and think this is a good reason to drop them. 🙂 I want to say somebody said that the year acts like a major release in some way, but I can't remember; otherwise, upper bounds make even less sense for CalVered projects. (FWIW I did include the upper bound on a PR for this that I abandoned, but it was more in line with existing Kedro standard.) > Also, are there any examples in the docs about the right

pip

syntax to install the different

kedro-dataset

optional datasets? Thanks! https://docs.kedro.org/en/stable/kedro_project_setup/dependencies.html#install-dependencies-at-a-type-level

@Deepyaman Datta thank you! I’d offer a suggestion: It’d be cool if we had those examples front and center at the

kedro-datasets

readme as well

Good idea! Here you go: https://github.com/kedro-org/kedro-plugins/pull/669 🙂