Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

Hello,
Can Polars dataframe be used as flyte task input &amp; output params out of the box? Since which version of flyte it’s supported.
I am using flyte 1.2.4.

Have you tried installing `flytekitplugins-polars` ? It has a type transformer for polars dataframes!

It exists for version 1.2.4 <https://pypi.org/project/flytekitplugins-polars/#history>!

Let me know if that isn't working for you

Here is a code snippet :slightly_smiling_face:
```# pip install flytekitplugins-polars

from flytekit import task, workflow
import polars as pl
import pandas as pd

@task
def t1() -&gt; pl.DataFrame:
    return pl.DataFrame(pd.DataFrame({"a": [1, 2, 3]}))

@task
def t2(df: pl.DataFrame) -&gt; int:
    return len(df)

@workflow
def wf() -&gt; int:
    df = t1()
    return t2(df=df)

wf()```

<@U03KECGDUBU>, thanks. I will try it out.
Looks like there is not much offered in flytekitplugins-polars yet, Therefore I will not install flytekitplugins-polars.

The plugin only adds type transformers so you can use them for inputs and outputs by themselves or as StructuredDatasets <@U044DJ094JG>.

I think having it support lazy execution would be a fantastic addition, but please let me know if you have other feature ideas for the plugin!