numpy tofile() with very large arrays saves all zeros

Question

When I try and save a very large (20000 x 20000 element) array, I get all zeros back:

In [2]: shape = (2e4,)*2

In [3]: r = np.random.randint(0, 10, shape)

In [4]: r.tofile('r.data')

In [5]: ls -lh r.data
-rw-r--r--  1 whg  staff   3.0G 23 Jul 16:18 r.data

In [6]: r[:6,:6]
Out[6]:
array([[6, 9, 8, 7, 4, 4],
       [5, 9, 5, 0, 9, 4],
       [6, 0, 9, 5, 7, 6],
       [4, 0, 8, 8, 4, 7],
       [8, 3, 3, 8, 7, 9],
       [5, 6, 1, 3, 1, 4]])

In [7]: r = np.fromfile('r.data', dtype=np.int64)

In [8]: r = r.reshape(shape)

In [9]: r[:6,:6]
Out[9]:
array([[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]])

np.save() does similar strange things.

After searching the net, I found that there is a known bug in OSX:

https://github.com/numpy/numpy/issues/2806

When I try to to read the the tostring() data from a file using Python's read(), I get a memory error.

Is there a better way of doing this? Can anyone recommend a pragmatic workaround to this problem?

user4815162342 · Accepted Answer · 2013-07-23 18:13:22Z

Use mmap to memory-map the file, and np.frombuffer to create an array that points into the buffer. Tested on x86_64 Linux:

# `r.data` created as in the question
>>> import mmap
>>> with open('r.data') as f:
...   m = mmap.mmap(f.fileno(), 0, mmap.MAP_SHARED, mmap.PROT_READ)
... 
>>> r = np.frombuffer(m, dtype='int64')
>>> r = r.reshape(shape)
>>> r[:6, :6]
array([[7, 5, 9, 5, 3, 5],
       [2, 7, 2, 6, 7, 0],
       [9, 4, 8, 2, 5, 0],
       [7, 2, 4, 6, 6, 7],
       [2, 9, 2, 2, 2, 6],
       [5, 2, 2, 6, 1, 5]])

Note that here r is a view of memory-mapped data, which makes it more memory-efficient, but comes with the side effect of automatically picking up changes to the file contents. If you want it to point to a private copy of the data, as the array returned by np.fromfile does, add an r = np.copy(r).

(Also, as written, this won't run under Windows, which requires slightly different mmap flags.)

Collectives™ on Stack Overflow

numpy tofile() with very large arrays saves all zeros

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related