I am not sure that this will work. The current else is covering the situation where data has type pd.DataFrame. We want to end up in a situation where we have a regular dataframe to work on. What does pd.DataFrame(pd.DataFrame(data)) actually do.

Also, if the int label is passed as part of a list or series then it will still be part of the resulting dataframe.

Can't we solve this by just reversing the order of the if clauses and leaving everything else untouched:

    if isinstance(data, pd.DataFrame):
        result = data
    elif isinstance(data, (list, pd.Series)):
        result = pd.DataFrame(data)
    else:
        # Handle other types (like str, int, float)
        result = pd.DataFrame([data])

Then the result.map(str) will apply to a dataframe containing a column that contains an int, which will convert it to a str.

And similarly reverse the if statements for extracting the result from the dataframe to return.

@girum-air you might also want to add a pipelines_tests.test_utils.PrepareLookupTestCase that tests each of these input types (str, int, list, Series, DataFrame)