SNOW-617426: Ability to overwrite tables when using write_pandas()

[iamontheinet]

What is the current behavior?

Using the write_pandas() API there is no way to overwrite contents of a Snowflake table with that of a Pandas DataFrame without first executing a truncate SQL statement.

What is the desired behavior?

Using the write_pandas() API there should be a way to optionally overwrite contents of a Snowflake table with that of a Pandas DataFrame.

How would this improve snowflake-connector-python?

Better user experience for Python developers and especially those devs that are not familiar with truncate SQL statement or know how to write SQL.

References, Other Background

Similar to overwriting contents of a Snowflake table when writing contents of Snowpark DataFrames using df.write.mode('overwrite')

[mhconradt]

Under the hood, write_pandas runs a COPY statement which is append-only. I recommend using TRUNCATE TABLE tbl IF EXISTS if you want to overwrite the rows in the table, or DROP TABLE IF EXISTS if you want to replace the table itself.

[iamontheinet]

Yes, I am looking into DROP TABLE IF EXIST which will account for any schema change(s) in the Pandas DataFrame. Thanks!

[simonwardjones]

I have made a similar issue here with a proposed solution: #1141

[simonwardjones]

@mhconradt would it be worth me making the PR I suggested or better to implement a work around in our application code?

[iamontheinet]

I am working on a solution as described in my comment above and will be creating a PR pretty soon.