This message was deleted.

yes - two options. # Option 1: do it in the driver.

from non_hamilton_script import historical_dataframe
import transforms # your hamilton functions

dr = driver.Driver({... your  config ...}, transforms)
df = dr.execute([ .... outputs ...], inputs={"historical_dataframe": historical_dataframe})

# Option 2: wrap import in a function with your transforms

# transforms.py

def historical_dataframe() -> pd.DataFrame:
    from non_hamilton_script import historical_dataframe
    return historical_dataframe  


def load_data_feed(path: str) -> pd.DataFrame:
    ...
    return df

def complete_dataset(load_data_feed: pd.DataFrame,
                     historical_dataframe: pd.DataFrame) -> pd.DataFrame:
    return pd.concat([historical_dataframe, load_data_feed])

Thank you. I like the second option since I can't seem to get the first to work. For the future I can see how the inputs argument would be extremely beneficial.

no, let me update the example

from non_hamilton_script import historical_dataframe
import transforms # your hamilton functions

dr = driver.Driver({... your  config ...}, transforms)
df = dr.execute([ .... outputs ...], inputs={"historical_dataframe": historical_dataframe()})

I believe the

historical_dataframe

needs to be instantiated in the inputs.

that’s a warning

yeah actually spyder just decided to break on both executions

I believe the

historical_dataframe

needs to be instantiated in the inputs.

yeah if there is a function to exercise, it needs to be run to create the dataframe so it’s passed in.

is there any issue with reusing arg name as the new function/column name as below?

# transforms.py
def historical_dataframe(historical_dataframe: pd.DataFrame) -> pd.DataFrame:
    return historical_dataframe.rename(columns=columns_mapping)

🤔 not sure about the kernel breaking

maybe this is what is actually breaking it

```def historical_dataframe(historical_dataframe_old: pd.DataFrame) -> pd.DataFrame:

return historical_dataframe.rename(columns=columns_mapping)```

just change the parameter name and then the input value accordingly.

yes this indirectly i guess fixed the error. option 1 work now.