Although IBM Netezza is not an integration tested database for turbodbc, some features have been added to support working with Netezza. Here are the recommended settings for connecting to a Netezza database via ODBC using the turbodbc module for Python.


Probably due to the prefetch buffering, turbodbc seems to be about as fast as using pyodbc, however, fetching using NumPy or Arrow calls (cursor.fetchallnumpy() or cursor.fetchallarrow()) is approximately 3 times faster due to avoiding the conversion from SQL types to Python types to NumPy/Arrow types and just converting directly to NumPy/Arrow types.