python - One hot encoding multi dimensional data -

using code below i'm attempting 1 hot code multi dimensional data. in case data 2d. code works expected 1d data 2d data each column 1 hot encoded instead of entire row. example : first data point [9,8] instead of single 1 hot encoded data point being generated 2 data points generated each corresponding 9 & 8 respectively.

how can 1 hot encode multi dimensional data ?

2dim.csv :

x_1,x_2 9,8 2,3 4,3 120,3 4,3 1,89 2,6 3,3 

code :

from sklearn.preprocessing import labelencoder keras.utils import np_utils import pandas pd  inputvalues = pd.read_csv('2dim.csv')  enc = inputvalues.apply(labelencoder().fit_transform) cat = np_utils.to_categorical(enc , 20)  inputvalues :  

cat :  

cat should contain 9 1 hot encoded data points instead of 18.

is possible solution combine every 2 adjoining inner arrays ? , example array([[1...n_1],[2...n_2],[3...n_3],[4...n_4]]) mapped array([[1...n_1,2...n_2],[3...n_3,4...n_4]])

try: pd.get_dummies(inputvalues) (docs)