Dagster Polars Bigquery IO manager does not store dates as date but datetime

[Loicqui]

import dagster as dg
import polars as pl
import datetime as dt

from HHM_dagster.lib.io_managers import IOManagerKeys


@dg.asset(
    key_prefix=["issues"],
    group_name="issues",
    partitions_def=dg.DailyPartitionsDefinition(
        start_date=dt.datetime(2024, 1, 1), timezone="Europe/Brussels", fmt="%Y-%m-%d"
    ),
    metadata={"partition_expr": "partition_key"},
    io_manager_key=IOManagerKeys.BIGQUERY_IO_MANAGER.value,
)
def my_upstream_partitioned_asset(context: dg.AssetExecutionContext) -> pl.DataFrame:
    # Your asset logic here
    return (
        pl.DataFrame({"partition_key": [context.partition_key]})
        .with_columns(pl.col("partition_key").str.to_date("%Y-%m-%d"))
        .with_columns(pl.col("partition_key").alias("date"))
    )

content of my resources :

IOManagerKeys.BIGQUERY_IO_MANAGER.value: PolarsBigQueryIOManager(
                project=settings.gcp_project.name,
                location=bigquery_zone,
                gcp_credentials=settings.gcp_project.bigquery_e64,
            )

When I launch the asset I get the following output metadata which confirms that the polars dataframe, before being loaded in bigquery has Date columns :

However in bigquery, the data is stored as timestamp

Why is that and how to solve this?
This is both for the partition column and the not partition column

[Loicqui]

I allow myself to up this issue as it still has the same effect