Closed
Description
Hi everybody,
I worked today with a lot of unicode data and deeply missed the option to define a default_charset when writing a dataframe to a new sql table. I did not know that pandas uses "latin-1" as default and ran into various strange encoding problems such as "Data too long for column" when inserting my unicode data. The exception was raised despite the explicit definition of utf-8 in the connection object.
As a workaround I created the table myself and let pandas only append the DataFrame. However, it would be much better if pandas could create the tables itself.
Best regards
Andy