不带（NaN）的pandas.DataFrame.merge

Question

我需要加入包含无/无的字段。但是熊猫出卖那是正常的吗？ （模拟相交，需要保留第一个dframe的索引）

a = pandas.DataFrame([[None, 1], [1, 2]], index=[10, 11], columns=['a', 'b'])
b = pandas.DataFrame([[None, 2], [1, 2]], columns=['c', 'd'])

a.merge(
    right=b.drop_duplicates().set_index(b.columns.to_list()),
    how='inner',
    left_on=a.columns.to_list(),
    right_index=True)

#       a  b
# 11  1.0  2
# 11  1.0  2 # Where does this row come from ???
#
# While i was expecting just:
#       a  b
# 11  1.0  2

但是

a = pandas.DataFrame([[None, 1], [1, 2]], index=[10, 11], columns=['a', 'b'])
b = pandas.DataFrame([[None, 1], [1, 2]], columns=['c', 'd'])

a.merge(
    right=b.drop_duplicates().set_index(b.columns.to_list()),
    how='inner',
    left_on=a.columns.to_list(),
    right_index=True)

#       a  b
# 11  1.0  2
# Where is the row [None, 1] ???
#
# While i was expecting:
#       a  b
# 10  NaN  1
# 11  1.0  2

如何与无人一起加入？