`polars.Time` not understood by Engine

[rhacs]

Describe the bug
Got the error TypeError: Data type 'Time' not understood by Engine. when I tried to use polars.Time as a Data Type on a DataFrameModel

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of pandera.
• (optional) I have confirmed this bug exists on the main branch of pandera.

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

from pandera import DataFrameModel, Field

from pandera.typing import Series

from pandera.engines.polars_engine import Time

# from polars import Time


class TimeDataModel(DataFrameModel):
    time: Series[Time] = Field(nullable=False, coerce=True)


if __name__ == "__main__":
    TimeDataModel.to_schema()

Expected behavior

Given that Time is inside the supported Data Types, it should work as intended like any other type.

Desktop (please complete the following information):

• OS: Windows 10
• Python: 3.12.9
• pandera: 0.23.1
• polars: 1.25.2

Additional info

Error when using pandera.engines.polars_engine.Time:

TypeError: Data type 'Time' not understood by Engine.

Error when using polars.Time:

TypeError: dtype 'Time' not understood

[ksolarski]

You need to import DataFrameModel from pandera.polars instead of pandera if your DataframeModel is using Polars data types or Polars engine. Editing your example, the following runs:

from pandera.polars import DataFrameModel, Field

from pandera.typing import Series

from pandera.engines.polars_engine import Time

# from polars import Time


class TimeDataModel(DataFrameModel):
    time: Series[Time] = Field(nullable=False, coerce=True)


if __name__ == "__main__":
    TimeDataModel.to_schema()

[rhacs]

@ksolarski thanks for replying

I was testing the workaround you gave me, but I couldn't make it work on my project, it was practically imploding, so I decided to try it independently.

This is the test file

from pandera.polars import DataFrameModel, Field
from pandera.typing import Series

from pandera.engines.polars_engine import Time

from pandas import DataFrame


class TestDataFrame(DataFrameModel):
    time: Series[Time] = Field(nullable=False, coerce=False)
    # time: Series[Time] = Field(nullable=False, coerce=True)


if __name__ == '__main__':
    test_data: DataFrame = DataFrame({
        "time": ["10:00", "11:00", "12:00", "13:00"]
    })

    TestDataFrame.validate(test_data, lazy=True)

And this is the error I get

Backend not found for backend, class: (<class 'pandera.api.polars.container.DataFrameSchema'>, <class 'pandas.core.frame.DataFrame'>). 
Looked up the following base classes: (<class 'pandas.core.frame.DataFrame'>, <class 'pandas.core.generic.NDFrame'>, <class 'pandas.core.base.PandasObject'>, <class 'pandas.core.accessor.DirNamesMixin'>, <class 'pandas.core.indexing.IndexingMixin'>, <class 'pandas.core.arraylike.OpsMixin'>, <class 'object'>)

Also tried to update the libraries in case I was missing something crucial, but the result is the same. I'm working with

• python 3.12.9
• pandera 0.23.1
• polars 1.25.2

[ksolarski]

You are trying to use time data type from Polars engine with Pandas dataframe. Pandas dataframe doesn't have support for time data type. You should use Polars instead:

from pandera.polars import DataFrameModel, Field
from pandera.typing import Series

from pandera.engines.polars_engine import Time
from datetime import time
from polars import DataFrame


class TestDataFrame(DataFrameModel):
    time: Series[Time] = Field(nullable=False, coerce=False)


test_data: DataFrame = DataFrame(
    {"time": [time(10, 00), time(11, 00), time(12, 00), time(13, 00)]},
)

TestDataFrame.validate(test_data, lazy=True)

[rhacs]

@ksolarski

That did the trick, thanks for the insight

I was using Pandas to read the file, gather the data and apply the validations using the data frame schema. Since everything was working so far, I thought the DataFrame from pandas was compatible with the one from polars

Thanks again, cheers

[cosmicBboy]

Closing issue, looks like this is addressed