Invalid Input Error

[neelasha23]

When running a select query without running the dataset download step I get error that file not found.

However, even after running the download step and rerunning I get the below error:

(I have repeated the steps by opening a clean notebook session and got the same set of errors)

[edublancas]

this isn't a bug per-se, but it'd be better if jupysql automatically performed the rollback.

we are already covering this for some PostgreSQL errors and we could handle this DuckDB edge case:

jupysql/src/sql/connection/connection.py

Line 836 in b538512

def _execute_with_error_handling(self, operation):

[bbeat2782]

Acceptance Criteria

1. Ensure downloading and rerunning the select statement works as expected
2. Add a test case for duckdb

[bbeat2782]

demo.mov

I tried to replicate the error, but it works as expected on my local setting.

EDIT: I was using sqlalchemy==1.4.49. After upgrading sqlalchemy version to 2.0.20, I was able to replicate the error