df['column_name'].iloc[0] # first element
df['column_name'].iloc[-1] # last element
#OR
df.column_name.head(1)
df.column_name.tail(1)
#OR if the column_one is sorted using sort_values(column_one)
df.nlargest(1, 'column_one')
df.nsmallest(1, 'column_one')