I have been converting polars dataframes to pandas for Dash AG tables, as shown here:
df_polars.to_pandas().to_dict('records')
The way to do this without converting to pandas is polars.to_dicts, as shown here:
df_polars.to_dicts()
Notice the added ‘s’, to_dict for pandas, to_dicts for polars. Also, polars.to_dicts() does not use ‘records’ as a part of the function call.