In [49]: 
from __future__ import division

import numpy as np
import matplotlib.pyplot as pt

import scipy.sparse as sps

Building a sparse matrix

In [50]: 
data = [5, 6, 7]
rows = [1, 1, 2]
columns = [2, 4, 6]

A = sps.coo_matrix(
        (data, (rows, columns)),
        shape=(10, 10), dtype=np.float64)
A
Out[50]:
<10x10 sparse matrix of type '<type 'numpy.float64'>'
	with 3 stored elements in COOrdinate format>
In [51]: 
A.todense()
Out[51]:
matrix([[ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],
        [ 0.,  0.,  5.,  0.,  6.,  0.,  0.,  0.,  0.,  0.],
        [ 0.,  0.,  0.,  0.,  0.,  0.,  7.,  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.,  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.,  0.]])
In [52]: 
A.nnz
Out[52]:
3
In [53]: 
pt.spy(A)
Out[53]:
<matplotlib.lines.Line2D at 0x333c8d0>

For a COO matrix, the juicy attributes are data, row, and col.

In []: 












Coordinate format is not the only format. For Compressed Sparse Row, look in data, indptr, and indices.









In [47]:

















How about performance?









In [43]:





fill_percent = 5

size = 1000
nentries = size**2 * fill_percent // 100

data = np.random.randn(nentries)
rows = (np.random.rand(nentries)*size).astype(np.int32)
columns = (np.random.rand(nentries)*size).astype(np.int32)

B_coo = sps.coo_matrix(
        (data, (rows, columns)),
        shape=(size, size), dtype=np.float64)

B_csr = sps.csr_matrix(B_coo)

B_dense = B_coo.todense()
















In [48]:





vec = np.random.randn(size)

from time import time
start = time()

for i in xrange(200):
    B_dense.dot(vec)
    
print "time: %g" % (time() - start)



















time: 0.282632





















In []: