How to show the full scikit's sparse matrix

Question

I am vectorizing with tfidf:

X = tfidf_vect.fit_transform(df['string'].values)

I would like to se the whole matrix of the above code so I tried this:

print X.toarray()

And obtained this:

[[0 0 0 ..., 0 0 0]
 [0 0 0 ..., 0 0 0]
 [0 0 0 ..., 0 0 0]
 ..., 
 [0 0 0 ..., 0 0 0]
 [0 0 0 ..., 0 0 0]
 [0 0 0 ..., 0 0 0]]

How can I write to some txt file or see the full matrix X?.

Do you want to see it with all the 0s, or would a list of non zero entries be better. What is the size (shape) of this array? — hpaulj, Feb 08 '15 at 04:49
There is an equivalent of the MATLAB `spy` function which plots an image of the sparsity of a sparse matrix: http://stackoverflow.com/questions/18651869/scipy-equivalent-for-matlab-spy — hpaulj, Feb 08 '15 at 08:11

score 1 · Accepted Answer · answered Feb 08 '15 at 05:15

1

Here is how you can write it to a text file:

pd.DataFrame(X.toarray()).to_csv('bow.csv')

Keep in mind that it can have very high 'n', and might make for a very large .txt

answered Feb 08 '15 at 05:15

JAB

score 1 · Answer 2 · answered Feb 08 '15 at 13:15

1

For interactive use, you can change the number of items to show

numpy.set_printoptions(edgeitems=1e9)

For saving to text files, use numpy.savetxt(X.toarray()) or some other similar function.

answered Feb 08 '15 at 13:15

pv.

2 Answers2