I have two dataframes:

import pandas as pd
import numpy as np

n=10

df = pd.DataFrame(
    {
        "A": 1.0,
        "B": pd.Timestamp("20130102"),
        "C": pd.Series(1, index=list(range(n)), dtype="float32"),
        "D": np.array([3] * n, dtype="int32"),
        "E": pd.Categorical(["test", "train"] * int(n/2)),
        "F": "foo",
    }
)

i_f = ['A', 'B']
o_f = ['C', 'D', 'E', 'F']
df1 = pd.DataFrame(data=df[i_f].values, columns=i_f).reset_index(drop=True)
df2 = pd.DataFrame(data=df[o_f].values, columns=o_f).reset_index(drop=True)

df1 and df2 are dataframes with the same number of rows, and different columns. Now I want to select a row i in df1, a row j in df2, and obtain a single dataframe of a single row, which has the columns from df1 and df2. I do:

i = 3
j = 2 
df_curr = pd.concat([df1.iloc[[i], :], df2.iloc[[j], :]], axis=1)

However, what I get instead is a dataframe with two rows, and with NaNs:

print(df_curr)
#      A          B    C    D     E    F
# 3  1.0 2013-01-02  NaN  NaN   NaN  NaN
# 2  NaN        NaT  1.0    3  test  foo

Instead, I need

       A          B    C    D     E    F
#    1.0 2013-01-02  1.0    3  test  foo

How can I get it?