Python Programming/numpy

Numpy is a numeric library for Python. Its main purpose is to provide fast manipulation of multidimensional arrays.

It's provided with the main Linux distribution, however it can be installed through the Debian package python-numpy. On systems with pip, running pip install numpy will work. On Windows, it can be downloaded on http://sourceforge.net/projects/numpy/files/.

Then, once the .zip unpacked, the installation is done by entering into the console:

python setup.py install

In case of error:

• ImportError “No Module named Setuptools”, save the file https://bootstrap.pypa.io/ez_setup.py into the Python folder, and install it with: python ez_setup.py install.
  • Microsoft Visual C++ 9.0 is required, download it on https://www.microsoft.com/en-us/download/details.aspx?id=44266

Arrays in numpy are of type numpy.ndarray. However, the most common way of defining them is using numpy.array, which is a function. For example,

import numpy
a = numpy.array([1, 2, 3])
b = numpy.array([[3, 1, 4, 1, 5],
                 [9, 2, 6, 5, 3]])

There are other ways to initialize numpy arrays. For example, to initialize a random array,

import numpy
rng = numpy.random.default_rng() # initialize random number generator
a = rng.random(5) # five random floats between 0 and 1
b = rng.random(3, 5) # a 3 by 5 array of floats between 0 and 1
c = rng.integers(0, 10, size=3) # 3 random integers between 0 and 10, excluding 10 and including 0

Some other ways of initialization are:

import numpy
a = numpy.zeros(2, 3) # 2 by 3 array of only zeros.
b = numpy.ones(3, 5) # 3 by 5 array of only ones.

import numpy
mydata = [numpy.random.normal(0,1) for i in range(10000) ]
h, n = numpy.histogram( mydata , 100, (-5,5) )

This section is a stub. You can help Wikibooks by expanding it.

• http://numpy.scipy.org/
• http://docs.scipy.org/doc/numpy/reference/
• Research using FOSS