SNOW-590657: Add overwrite option to write_pandas

[simonwardjones]

What is the current behavior?

When using write_pandas method with auto_create_table as True the code generates a create statement by inferring the column types from the staged dataframe (parquet) file.

The code snippet (from pandas_tools.py) that does this is here:

create_table_sql = (
    f"CREATE {'TEMP ' if create_temp_table else ''}TABLE IF NOT EXISTS {location} "
    f"({create_table_columns})"
    f" /* Python:snowflake.connector.pandas_tools.write_pandas() */ "
)

What is the desired behavior?

It would be handy if we could add an extra argument to the function overwrite that would enable this to be a create or replace with something like:

create_table_sql = (
    f"CREATE {'OR REPLACE' if overwrite else ''} "
    f"{'TEMP ' if create_temp_table else ''} TABLE"
    f"{'IF NOT EXISTS' if not overwrite else ''} {location} "
    f"({create_table_columns})"
    f" /* Python:snowflake.connector.pandas_tools.write_pandas() */ "
)

How would this improve snowflake-connector-python?

It would enable users to overwrite a table with a dataframe (with potentially different shape and column definitions) without having to run an explicit drop statement first.

References, Other Background

[simonwardjones]

Just a though if this were to be added there might want to be a validation that auto_create_table is not False and overwrite = True

[simonwardjones]

I am happy to raise a PR but just wanted to get some feedback first