CODEWITHSUNDEEP
pythonnumpy
silent_grave
silent_grave

Reputation: 638

NumPy: Error While Concatenation - zero-dimensional arrays cannot be concatenated

I am trying to concat two valid array via np.concat() method.

My code:

print X_train.shape, train_names.shape
X_train = np.concatenate([train_names,X_train], axis=0)

The output:



    (3545, 93355) (3545, 692)


    ValueError                                Traceback (most recent call last)
    <ipython-input-58-59dc66874663> in <module>()
      1 print X_train.shape, train_names.shape
    ----> 2 X_train = np.concatenate([train_names,X_train], axis=0)
      

    ValueError: zero-dimensional arrays cannot be concatenated

As you can see, the shapes of arrays align, still I am getting this weird error. Why?

EDIT: I have tried with axis=1 as well. Same result

EDIT 2: Eqauted data types using .astype(np.float64). Same result.

Upvotes: 6

Views: 11472

Answers (3)

Ashish Gupta
Ashish Gupta

Reputation: 185

In my case, the problem was due to sparse nature of matrix returned by LabelEncoder()

so this line gave me an error:

np.append(airlineTrain, train_transformed,axis =1)

To fix it, I used this: 

np.append(airlineTrain.toarray(), train_transformed.toarray(),axis =1 )

Alternately, you may be using NLTK, where the sparse storage of matrices can be converted by using todense()

Upvotes: 1

hpaulj
hpaulj

Reputation: 231395

Applying np.concatenate to scipy sparse matrices produces this error:

In [162]: from scipy import sparse
In [163]: x=sparse.eye(3)
In [164]: x
Out[164]: 
<3x3 sparse matrix of type '<class 'numpy.float64'>'
    with 3 stored elements (1 diagonals) in DIAgonal format>
In [165]: np.concatenate((x,x))
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-165-0b67d0029ca6> in <module>()
----> 1 np.concatenate((x,x))

ValueError: zero-dimensional arrays cannot be concatenated

There are sparse functions to do this:

In [168]: sparse.hstack((x,x)).A
Out[168]: 
array([[ 1.,  0.,  0.,  1.,  0.,  0.],
       [ 0.,  1.,  0.,  0.,  1.,  0.],
       [ 0.,  0.,  1.,  0.,  0.,  1.]])
In [169]: sparse.vstack((x,x)).A
Out[169]: 
array([[ 1.,  0.,  0.],
       [ 0.,  1.,  0.],
       [ 0.,  0.,  1.],
       [ 1.,  0.,  0.],
       [ 0.,  1.,  0.],
       [ 0.,  0.,  1.]])

Upvotes: 15

xthestreams
xthestreams

Reputation: 169

Pass the arrays as a tuple rather than a list. X_train = np.concatenate((train_names,X_train), axis=0)

Upvotes: 1

Related Questions