Pandas DataFrames have an itertuples method that returns a generator of named tuples. The named tuples have an _asdict() method. So you can get a generator of rows as dictionaries with:

row_generator = (row._asdict() for row in df.itertuples())

References:

• df.itertuples: https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.itertuples.html#pandas.DataFrame.itertuples
• namedtuple._asdict(): https://docs.python.org/3/library/collections.html#collections.somenamedtuple._asdict