CODEWITHSUNDEEP
pythonpandas
Allan Tanaka
Allan Tanaka

Reputation: 307

How to unconcat pandas dataframe?

I want to use some function that requires to concat two dataframe. Here is the example:

import numpy as np
import pandas as pd
data1 = np.array([['','Col1','Col2'],['1',1,2],['2',3,4]])
data1=pd.DataFrame(data=data1[1:,1:],index=data1[1:,0],columns=data1[0,1:])
data2=np.array([['','Col1','Col2'],['1',5,6],['2',7,8]])
data2=pd.DataFrame(data=data2[1:,1:],index=data2[1:,0],columns=data2[0,1:])
X=pd.concat([data1,data2],0)
X_transformed=func(X)

Now I want to unconcat X_transformed back into the original data1 and data2. Is there a way to do so?

Upvotes: 1

Views: 274

Answers (2)

jezrael
jezrael

Reputation: 862581

You can add parameter keys to concat for distinguish each DataFrame and then select by loc:

X=pd.concat([data1,data2],0, keys=[0,1])
print (X)
    Col1 Col2
0 1    1    2
  2    3    4
1 1    5    6
  2    7    8

data11 = X.loc[0]
data22 = X.loc[1]
print (data11)
  Col1 Col2
1    1    2
2    3    4

print (data22)
  Col1 Col2
1    5    6
2    7    8

EDIT:

More general solution:

data3=data2.iloc[[0]].rename({'1':'10'})

dfs = [data1,data2,data3]
X=pd.concat(dfs, keys=np.arange(len(dfs)))
print (X)
     Col1 Col2
0 1     1    2
  2     3    4
1 1     5    6
  2     7    8
2 10    5    6

print (X.xs(0))
  Col1 Col2
1    1    2
2    3    4

print (X.xs(1))
  Col1 Col2
1    5    6
2    7    8

print (X.xs(2))
   Col1 Col2
10    5    6

Upvotes: 2

DYZ
DYZ

Reputation: 57033

X_transformed.iloc[:data1.shape[0]] is the first dataframe. X_transformed.iloc[data1.shape[0]:] is the second one.

Upvotes: 0

Related Questions