df_grouped = df.groupby('column')

​

# iterate over each group

for group_name, df_group in df_grouped:

    # iterate over each row

    for row_index, row in df_group.iterrows():

      # do your stuff