vdk-core: KeyError in user code coming from pandas classified as PlatformError

[DeltaMichael]

Steps To Reproduce

Run a job in vdk with the following step

import logging
import pandas as pd

from vdk.api.job_input import IJobInput

log = logging.getLogger(__name__)

def run(job_input: IJobInput):
    """
    Function named `run` is required in order for a python script to be recognized as a Data Job Python step and executed.

    VDK provides to every python step an object - job_input - that has methods for:

    * executing queries to OLAP Database;
    * ingesting data into a database;
    * processing data into a database.
    See IJobInput documentation for more details.
    """
    log.info(f"Starting job step {__name__}")
    df1 = pd.DataFrame({'lkey': ['foo', 'bar', 'baz', 'foo'],
                    'value': [1, 2, 3, 5]})
    df2 = pd.DataFrame({'rkey': ['foo', 'bar', 'baz', 'foo'],
                    'value': [5, 6, 7, 8]})
    # df1.merge(df2, left_on='lkey', right_on='rkey')
    result = pd.merge(df1, df2, how="inner", on=["timestamp", "status_code", "response_status", "exception", "region", "region_type", "tenant", "utc_time"])

Expected behavior
KeyError is thrown and job ends with UserError.

Actual behavior
KeyError is thrown and job ends with PlatformError