In [1]:

    

import numpy as np
from scipy.sparse import csr_matrix, csc_matrix
from scipy.sparse.linalg import inv, eigs


    

In [2]:

    

l = [[2, 5],
     [1, 3]]


    

In [3]:

    

csr = csr_matrix(l)
csc = csc_matrix(l)


    

In [4]:

    

# csr_inv = linalg.inv(csr)
# SparseEfficiencyWarning: splu requires CSC matrix format
# SparseEfficiencyWarning: spsolve is more efficient when sparse b is in the CSC matrix format


    

In [5]:

    

csc_inv = inv(csc)
print(csc_inv)


    

    




  (0, 0)	3.0
  (1, 0)	-1.0
  (0, 1)	-5.0
  (1, 1)	2.0

In [6]:

    

print(type(csc_inv))


    

    




<class 'scipy.sparse.csc.csc_matrix'>

In [7]:

    

print(csc_inv.toarray())


    

    




[[ 3. -5.]
 [-1.  2.]]

In [8]:

    

l = [[1, 1, 2],
     [0, 2, -1],
     [0, 0, 3]]


    

In [9]:

    

w, v = np.linalg.eig(l)


    

In [10]:

    

print(w)


    

    




[1. 2. 3.]

In [11]:

    

print(v)


    

    




[[ 1.          0.70710678  0.33333333]
 [ 0.          0.70710678 -0.66666667]
 [ 0.          0.          0.66666667]]

In [12]:

    

csr_f = csr_matrix(l, dtype=float)
csr_i = csc_matrix(l, dtype=int)


    

In [13]:

    

w, v = eigs(csr_f, k=1)


    

In [14]:

    

print(w)


    

    




[3.+0.j]

In [15]:

    

print(v)


    

    




[[ 0.33333333+0.j]
 [-0.66666667+0.j]
 [ 0.66666667+0.j]]

In [16]:

    

# w, v = eigs(csr_f, k=2)
# TypeError: Cannot use scipy.linalg.eig for sparse A with k >= N - 1. Use scipy.linalg.eig(A.toarray()) or reduce k.


    

In [17]:

    

# w, v = eigs(csr_i, k=1)
# ValueError: matrix type must be 'f', 'd', 'F', or 'D'