SNOW-1552772: FutureWarning from Pandas about indexing in mock._pandas_util

[NickFishr]

1. What version of Python are you using?

   Python 3.11.0rc1 (main, Aug 12 2022, 10:02:14) [GCC 11.2.0]
   

2. What are the Snowpark Python and pandas versions in the environment?

   pandas==2.2.1
   snowflake-snowpark-python==1.20.0

3. What did you do?

   Created a local testing session and tried to create some Snowpark dataframes from Pandas ones.

   import pandas as pd
   from snowflake.snowpark import Session
   
   session = Session.builder.config("local_testing", True).create()
   
   pandas_df = pd.DataFrame({"Timestamp": [pd.to_datetime(1490195805, unit="s")]})
   df = session.create_dataframe(pandas_df)
   
   pandas_df = pd.DataFrame([[1.0]], columns=["A"], dtype=pd.Float64Dtype())
   df = session.create_dataframe(pandas_df)

4. What did you expect to see?

   No warnings.

   Instead Pandas threw the following warnings:

   /tmp/snowpark-python/src/snowflake/snowpark/mock/_pandas_util.py:96: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
     if isinstance(data.dtypes[col_idx], pd.DatetimeTZDtype):
   /tmp/snowpark-python/src/snowflake/snowpark/mock/_functions.py:811: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
     data = row[0]
   /tmp/snowpark-python/src/snowflake/snowpark/mock/_pandas_util.py:82: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
     if isinstance(data.dtypes[col_idx], PANDAS_INTEGER_TYPES)
   /tmp/snowpark-python/src/snowflake/snowpark/mock/_pandas_util.py:83: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
     else float(str(data.iloc[row_idx][col_idx]))
   

Additional Notes

This is possibly not a real bug and could've been muted with a warnings.simplefilter (like here).

But it is a FutureWarning from Pandas so it will probably come back later to bite us and it seems easy to fix.

[sfc-gh-sghosh]

Hello @NickFishr ,

Thank you for raising the issue, yes, we can see the warning with local_testing. There is no issue with regular sessions.
Will update.

Regards,
Sujan

[sfc-gh-aling]

thanks for the PR! we will merge once the tests pass

[sfc-gh-aling]

PR merged, it will be carried in our next release, thanks again!