I have two DataFrames with different indexes like:
import pandas as pd
a = pd.DataFrame([1, 2, 3], index=['a', 'b', 'c'],
columns=['one'])
b = pd.DataFrame([5, 6], index=['d', 'e'],
columns=['two'])
And i can create "Cartesian" MultiIndex using:
a_plus_b = pd.MultiIndex.from_product([a.index,b.index])
Which turns to an empty MultiIndex:
MultiIndex(levels=[['a', 'b', 'c'], ['d', 'e']],
labels=[[0, 0, 1, 1, 2, 2], [0, 1, 0, 1, 0, 1]])
How to create Cartesian sum like the following?
'a' 'd' 6 # 1 + 5
'e' 7 # 1 + 6
'b' 'd' 7 # 2 + 5
'e' 8 # 2 + 6
'c' 'd' 8 # 3 + 5
'e' 9 # 3 + 6
Use reindex
by first and second level:
s = a['one'].reindex(a_plus_b, level=0) + b['two'].reindex(a_plus_b, level=1)
print (s)
a d 6
e 7
b d 7
e 8
c d 8
e 9
dtype: int64
You can avoid the intermediary step of creating a MultiIndex
explicitly by using pd.merge
:
res = pd.merge(a.rename_axis('A').reset_index().assign(key=1),
b.rename_axis('B').reset_index().assign(key=1), on='key')
res = res.assign(total=res['one'] + res['two'])\
.groupby(['A', 'B'])['total'].sum()
print(res)
A B
a d 6
e 7
b d 7
e 8
c d 8
e 9
Name: total, dtype: int64
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.