Polars equivalent of pandas.ExcelDataset

[wolthom]

Description

For pandas dataframes, a specific output dataset for Excel files is supported (Documentation https://docs.kedro.org/projects/kedro-datasets/en/kedro-datasets-6.0.0/api/kedro_datasets.pandas.ExcelDataset.html)
One very appealing feature is the ability to pass in a dictionary of dataframes and generate a multi-worksheet output file.

Context

For the kedro pipelines I am implementing at work, I often implement debug nodes alongside the production pipelines.
Said debug nodes typically output Excel files or visualizations so they are easy to review and interactively browse.
It's extremely useful to be able to create various pre-aggregations, filtrations etc. and output them all in one spreadsheet I can share e.g. with Stakeholders.

Possible Implementation

I don't know the kedro / kedro-datasets architecture well enough yet to propose an implementation strategy.

Possible Alternatives

My current workaround is to simply do my implementations in polars, but transform to pandas dataframes for the output.
While not a huge issue, I would like to keep my dependency tree as small as possible and do not actually use pandas anymore.

[Stephanieewelu]

Hi @wolthom , I'd love to contribute to this issue! My plan is to create a PolarsExcelDataset that mimics the existing pandas.ExcelDataset functionality, including support for multi-sheet output. I'll follow the Kedro contribution guidelines and submit a PR soon. Let me know if you have any specific requirements. Thanks!

[ankatiyar]

Hi @wolthom and @Stephanieewelu,
In kedro-datasets we currently have the polars.EagerDataset and polars.LazyDataset which can also work with Excel files. Is this useful for what you're trying to do?

Here's the contributing guide to kedro-datasets https://github.com/kedro-org/kedro-plugins/blob/main/kedro-datasets/CONTRIBUTING.md