NumPy: Beginner's Guide - Third Edition

By Ivan Idris
    What do you get with a Packt Subscription?

  • Instant access to this title and 7,500+ eBooks & Videos
  • Constantly updated with 100+ new titles each month
  • Breadth and depth in over 1,000+ technologies
  1. Free Chapter
    NumPy Quick Start
About this book

In today's world of science and technology, it's all about speed and flexibility. When it comes to scientific computing, NumPy tops the list. NumPy will give you both speed and high productivity. This book will walk you through NumPy with clear, step-by-step examples and just the right amount of theory. The book focuses on the fundamentals of NumPy, including array objects, functions, and matrices, each of them explained with practical examples. You will then learn about different NumPy modules while performing mathematical operations such as calculating the Fourier transform, finding the inverse of a matrix, and determining eigenvalues, among many others. This book is a one-stop solution to knowing the ins and outs of the vast NumPy library, empowering you to use its wide range of mathematical features to build efficient, high-speed programs.

Publication date:
June 2015


Chapter 1. NumPy Quick Start

Let's get started. We will install NumPy and related software on different operating systems and have a look at some simple code that uses NumPy. This chapter briefly introduces the IPython interactive shell. SciPy is closely related to NumPy, so you will see the SciPy name appearing here and there. At the end of this chapter, you will find pointers on how to find additional information online if you get stuck or are uncertain about the best way to solve problems.

In this chapter, you will cover the following topics:

  • Install Python, SciPy, matplotlib, IPython, and NumPy on Windows, Linux, and Macintosh

  • Do a short refresher of Python

  • Write simple NumPy code

  • Get to know IPython

  • Browse online documentation and resources



NumPy is based on Python, so you need to have Python installed. On some operating systems, Python is already installed. However, you need to check whether the Python version corresponds with the NumPy version you want to install. There are many implementations of Python, including commercial implementations and distributions. In this book, we focus on the standard CPython implementation, which is guaranteed to be compatible with NumPy.


Time for action – installing Python on different operating systems

NumPy has binary installers for Windows, various Linux distributions, and Mac OS X at There is also a source distribution, if you prefer that. You need to have Python 2.4.x or above installed on your system. We will go through the various steps required to install Python on the following operating systems:

  • Debian and Ubuntu: Python might already be installed on Debian and Ubuntu, but the development headers are usually not. On Debian and Ubuntu, install the python and python-dev packages with the following commands:

    $ [sudo] apt-get install python
    $ [sudo] apt-get install python-dev
  • Windows: The Windows Python installer is available at On this website, we can also find installers for Mac OS X and source archives for Linux, UNIX, and Mac OS X.

  • Mac: Python comes preinstalled on Mac OS X. We can also get Python through MacPorts, Fink, Homebrew, or similar projects.

    Install, for instance, the Python 2.7 port by running the following command:

    $ [sudo] port install python27

    Linear Algebra PACKage (LAPACK) does not need to be present but, if it is, NumPy will detect it and use it during the installation phase. It is recommended that you install LAPACK for serious numerical analysis as it has useful numerical linear algebra functionality.

What just happened?

We installed Python on Debian, Ubuntu, Windows, and the Mac OS X.


You can download the example code files for all the Packt books you have purchased from your account at If you purchased this book elsewhere, you can visit and register to have the files e-mailed directly to you.


The Python help system

Before we start the NumPy introduction, let's take a brief tour of the Python help system, in case you have forgotten how it works or are not very familiar with it. The Python help system allows you to look up documentation from the interactive Python shell. A shell is an interactive program, which accepts commands and executes them for you.


Time for action – using the Python help system

Depending on your operating system, you can access the Python shell with special applications, usually a terminal of some sort.

  1. In such a terminal, type the following command to start a Python shell:

    $ python
  2. You will get a short message with the Python version and other information and the following prompt:


    Type the following in the prompt:

    >>> help()

    Another message appears and the prompt changes as follows:

  3. If you type, for instance, keywords as the message says, you get a list of keywords. The topics command gives a list of topics. If you type any of the topic names (such as LISTS) in the prompt, you get additional information about the topic. Typing q quits the information screen. Pressing Ctrl + D together returns you to the normal Python prompt:


    Pressing Ctrl + D together again ends the Python shell session.

What just happened?

We learned about the Python interactive shell and the Python help system.


Basic arithmetic and variable assignment

In the Time for action – using the Python help system section, we used the Python shell to look up documentation. We can also use Python as a calculator. By the way, this is just a refresher, so if you are completely new to Python, I recommend taking some time to learn the basics. If you put your mind to it, learning basic Python should not take you more than a couple of weeks.


Time for action – using Python as a calculator

We can use Python as a calculator as follows:

  1. In a Python shell, add 2 and 2 as follows:

    >>> 2 + 2
  2. Multiply 2 and 2 as follows:

    >>> 2 * 2
  3. Divide 2 and 2 as follows:

    >>> 2/2
  4. If you have programmed before, you probably know that dividing is a bit tricky since there are different types of dividing. For a calculator, the result is usually adequate, but the following division may not be what you were expecting:

    >>> 3/2

    We will discuss what this result is about in several later chapters of this book. Take the cube of 2 as follows:

    >>> 2 ** 3

What just happened?

We used the Python shell as a calculator and performed addition, multiplication, division, and exponentiation.


Time for action – assigning values to variables

Assigning values to variables in Python works in a similar way to most programming languages.

  1. For instance, assign the value of 2 to a variable named var as follows:

    >>> var = 2
    >>> var
  2. We defined the variable and assigned it a value. In this Python code, the type of the variable is not fixed. We can make the variable in to a list, which is a built-in Python type corresponding to an ordered sequence of values. Assign a list to var as follows:

    >>> var = [2, 'spam', 'eggs']
    >>> var
    [2, 'spam', 'eggs']

    We can assign a new value to a list item using its index number (counting starts from 0). Assign a new value to the first list element:

    >>> var
    ['ham', 'spam', 'eggs']
  3. We can also swap values easily. Define two variables and swap their values:

    >>> a = 1
    >>> b = 2
    >>> a, b = b, a
    >>> a
    >>> b

What just happened?

We assigned values to variables and Python list items. This section is by no means exhaustive; therefore, if you are struggling, please read Appendix B, Additional Online Resources, to find recommended Python tutorials.


The print() function

If you haven't programmed in Python for a while or are a Python novice, you may be confused about the Python 2 versus Python 3 discussions. In a nutshell, the latest version Python 3 is not backward compatible with the older Python 2 because the Python development team felt that some issues were fundamental and therefore warranted a radical change. The Python team has committed to maintain Python 2 until 2020. This may be problematic for the people who still depend on Python 2 in some way. The consequence for the print() function is that we have two types of syntax.


Time for action – printing with the print() function

We can print using the print() function as follows:

  1. The old syntax is as follows:

    >>> print 'Hello'
  2. The new Python 3 syntax is as follows:

    >>> print('Hello')

    The parentheses are now mandatory in Python 3. In this book, I try to use the new syntax as much as possible; however, I use Python 2 to be on the safe side. To enforce the syntax, each Python 2 script with print() calls in this book starts with:

    >>> from __future__ import print_function
  3. Try to use the old syntax to get the following error message:

    >>> print 'Hello'
      File "<stdin>", line 1
        print 'Hello'
    SyntaxError: invalid syntax
  4. To print a newline, use the following syntax:

    >>> print()
  5. To print multiple items, separate them with commas:

    >>> print(2, 'ham', 'egg')
    2 ham egg
  6. By default, Python separates the printed values with spaces and prints output to the screen. You can customize these settings. Read more about this function by typing the following command:

    >>> help(print)

    You can exit again by typing q.

What just happened?

We learned about the print() function and its relation to Python 2 and Python 3.


Code comments

Commenting code is a best practice with the goal of making code clearer for yourself and other coders (see Usually, companies and other organizations have policies regarding code comment such as comment templates. In this book, I did not comment the code in such a fashion for brevity and because the text in the book should clarify the code.


Time for action – commenting code

The most basic comment starts with a hash sign and continues until the end of the line:

  1. Comment code with this type of comment as follows:

    >>> # Comment from hash to end of line
  2. However, if the hash sign is between single or double quotes, then we have a string, which is an ordered sequence of characters:

    >>> astring = '# This is not a comment'
    >>> astring
    '# This is not a comment'
  3. We can also comment multiple lines as a block. This is useful if you want to write a more detailed description of the code. Comment multiple lines as follows:

     Chapter 1 of NumPy Beginners Guide.
     Another line of comment.

    We refer to this type of comment as triple-quoted for obvious reasons. It also is used to test code. You can read about testing in Chapter 8, Assuring Quality with Testing.


The if statement

The if statement in Python has a bit different syntax to other languages, such as C++ and Java. The most important difference is that indentation matters, which I hope you are aware of.


Time for action – deciding with the if statement

We can use the if statement in the following ways:

  1. Check whether a number is negative as follows:

    >>> if 42 < 0:
    ...     print('Negative')
    ... else:
    ...     print('Not negative')
    Not negative

    In the preceding example, Python decided that 42 is not negative. The else clause is optional. The comparison operators are equivalent to the ones in C++, Java, and similar languages.

  2. Python also has a chained branching logic compound statement for multiple tests similar to the switch statement in C++, Java, and other programming languages. Decide whether a number is negative, 0, or positive as follows:

    >>> a = -42
    >>> if a < 0:
    ...     print('Negative')
    ... elif a == 0:
    ...     print('Zero')
    ... else:
    ...     print('Positive')

    This time, Python decided that 42 is negative.

What just happened?

We learned how to do branching logic in Python.


The for loop

Python has a for statement with the same purpose as the equivalent construct in C++, Pascal, Java, and other languages. However, the mechanism of looping is a bit different.


Time for action – repeating instructions with loops

We can use the for loop in the following ways:

  1. Loop over an ordered sequence, such as a list, and print each item as follows:

    >>> food = ['ham', 'egg', 'spam']
    >>> for snack in food:
    ...     print(snack)
  2. And remember that, as always, indentation matters in Python. We loop over a range of values with the built-in range() or xrange() functions. The latter function is slightly more efficient in certain cases. Loop over the numbers 1-9 with a step of 2 as follows:

    >>> for i in range(1, 9, 2):
    ...     print(i)
  3. The start and step parameter of the range() function are optional with default values of 1. We can also prematurely end a loop. Loop over the numbers 0-9 and break out of the loop when you reach 3:

    >>> for i in range(9):
    ...     print(i)
    ...     if i == 3:
    ...     print('Three')
    ...     break
  4. The loop stopped at 3 and we did not print the higher numbers. Instead of leaving the loop, we can also get out of the current iteration. Print the numbers 0-4, skipping 3 as follows:

    >>> for i in range(5):
    ...     if i == 3:
    ...             print('Three')
    ...             continue
    ...     print(i)
  5. The last line in the loop was not executed when we reached 3 because of the continue statement. In Python, the for loop can have an else statement attached to it. Add an else clause as follows:

    >>> for i in range(5):
    ...     print(i)
    ... else:
    ...     print(i, 'in else clause')
    (4, 'in else clause')
  6. Python executes the code in the else clause last. Python also has a while loop. I do not use it that much because the for loop is more useful in my opinion.

What just happened?

We learned how to repeat instructions in Python with loops. This section included the break and continue statements, which exit and continue looping.


Python functions

Functions are callable blocks of code. We call functions by the name we give them.


Time for action – defining functions

Let's define the following simple function:

  1. Print Hello and a given name in the following way:

    >>> def print_hello(name):
    ...     print('Hello ' + name)

    Call the function as follows:

    >>> print_hello('Ivan')
    Hello Ivan
  2. Some functions do not have arguments, or the arguments have default values. Give the function a default argument value as follows:

    >>> def print_hello(name='Ivan'):
    ...     print('Hello ' + name)
    >>> print_hello()
    Hello Ivan
  3. Usually, we want to return a value. Define a function, which doubles input values as follows:

    >>> def double(number):
    ...     return 2 * number
    >>> double(3)

What just happened?

We learned how to define functions. Functions can have default argument values and return values.


Python modules

A file containing Python code is called a module. A module can import other modules, functions in other modules, and other parts of modules. The filenames of Python modules end with .py. The name of the module is the same as the filename minus the .py suffix.


Time for action – importing modules

Importing modules can be done in the following manner:

  1. If the filename is, for instance,, import it as follows:

    >>> import mymodule
  2. The standard Python distribution has a math module. After importing it, list the functions and attributes in the module as follows:

    >>> import math
    >>> dir(math)
    ['__doc__', '__file__', '__name__', '__package__', 'acos', 'acosh', 'asin', 'asinh', 'atan', 'atan2', 'atanh', 'ceil', 'copysign', 'cos', 'cosh', 'degrees', 'e', 'erf', 'erfc', 'exp', 'expm1', 'fabs', 'factorial', 'floor', 'fmod', 'frexp', 'fsum', 'gamma', 'hypot', 'isinf', 'isnan', 'ldexp', 'lgamma', 'log', 'log10', 'log1p', 'modf', 'pi', 'pow', 'radians', 'sin', 'sinh', 'sqrt', 'tan', 'tanh', 'trunc']
  3. Call the pow() function in the math module:

    >>> math.pow(2, 3)

    Notice the dot in the syntax. We can also import a function directly and call it by its short name. Import and call the pow() function as follows:

    >>> from math import pow
    >>> pow(2, 3)
  4. Python lets us define aliases for imported modules and functions. This is a good time to introduce the import conventions we are going to use for NumPy and a plotting library we will use a lot:

    import numpy as np
    import matplotlib.pyplot as plt

What just happened?

We learned about modules, importing modules, importing functions, calling functions in modules, and the import conventions of this book. This concludes the Python refresher.


NumPy on Windows

Installing NumPy on Windows is straightforward. You only need to download an installer, and a wizard will guide you through the installation steps.


Time for action – installing NumPy, matplotlib, SciPy, and IPython on Windows

Installing NumPy on Windows is necessary but this is, fortunately, a straightforward task that we will cover in detail. It is recommended that you install matplotlib, SciPy, and IPython. However, this is not required to enjoy this book. The actions we will take are as follows:

  1. Download a NumPy installer for Windows from the SourceForge website

    Choose the appropriate NumPy version according to your Python version. In the preceding screen shot, we chose numpy-1.9.2-win32-superpack-python2.7.exe.

  2. Open the EXE installer by double-clicking on it as shown in the following screen shot:

  3. Now, we can see a description of NumPy and its features. Click on Next.

  4. If you have Python installed, it should automatically be detected. If it is not detected, your path settings might be wrong. At the end of this chapter, we have listed resources in case you have problems with installing NumPy.

  5. In this example, Python 2.7 was found. Click on Next if Python is found; otherwise, click on Cancel and install Python (NumPy cannot be installed without Python). Click on Next. This is the point of no return. Well, kind of, but it is best to make sure that you are installing to the proper directory and so on and so forth. Now the real installation starts. This may take a while.

    Install SciPy and matplotlib with the Enthought Canopy distribution ( It might be necessary to put the msvcp71.dll file in your C:\Windows\system32 directory. You can get it from A Windows IPython installer is available on the IPython website (see

What just happened?

We installed NumPy, SciPy, matplotlib, and IPython on Windows.


NumPy on Linux

Installing NumPy and its related recommended software on Linux depends on the distribution you have. We will discuss how you will install NumPy from the command line, although you can probably use graphical installers; it depends on your distribution (distro). The commands to install matplotlib, SciPy, and IPython are the same—only the package names are different. Installing matplotlib, SciPy, and IPython is recommended, but optional.


Time for action – installing NumPy, matplotlib, SciPy, and IPython on Linux

Most Linux distributions have NumPy packages. We will go through the necessary commands for some of the most popular Linux distros:

  • Installing NumPy on Red Hat: Run the following instructions from the command line:

    $ yum install python-numpy
  • Installing NumPy on Mandriva: To install NumPy on Mandriva, run the following command line instruction:

    $ urpmi python-numpy
  • Installing NumPy on Gentoo: To install NumPy on Gentoo, run the following command line instruction:

    $ [sudo] emerge numpy
  • Installing NumPy on Debian and Ubuntu: On Debian or Ubuntu, type the following on the command line:

    $ [sudo] apt-get install python-numpy

    The following table gives an overview of the Linux distributions and the corresponding package names for NumPy, SciPy, matplotlib, and IPython:

    Linux distribution





    Arch Linux





















    python-numpy, python-numpy-devel










NumPy on Mac OS X

You can install NumPy, matplotlib, and SciPy on the Mac OS X with a GUI installer (not possible for all versions) or from the command line with a port manager such as MacPorts, Homebrew, or Fink, depending on your preference. You can also install using a script from


Time for action – installing NumPy, SciPy, matplotlib, and IPython with MacPorts or Fink

Alternatively, we can install NumPy, SciPy, matplotlib, and IPython through the MacPorts route or with Fink. The following installation steps show how to install all these packages:

  • Installing with MacPorts: Type the following command:

    $ [sudo] port install py-numpy py-scipy py-matplotlib py-ipython
  • Installing with Fink: Fink also has packages for NumPy—scipy-core-py24, scipy-core-py25, and scipy-core-py26. The SciPy packages are scipy-py24, scipy-py25 and scipy-py26. We can install NumPy and the additional recommended packages, referring to this book on Python 2.7, using the following command:

    $ fink install scipy-core-py27 scipy-py27 matplotlib-py27

What just happened?

We installed NumPy and the additional recommended software on Mac OS X with MacPorts and Fink.


Building from source

We can retrieve the source code for NumPy with git as follows:

$ git clone git:// numpy

Alternatively, download the source from

Install in /usr/local with the following command:

$ python build
$ [sudo] python install --prefix=/usr/local

To build, we need a C compiler such as GCC and the Python header files in the python-dev or python-devel packages.



After going through the installation of NumPy, it's time to have a look at NumPy arrays. NumPy arrays are more efficient than Python lists when it comes to numerical operations. NumPy code requires less explicit loops than the equivalent Python code.


Time for action – adding vectors

Imagine that we want to add two vectors called a and b (see Vector is used here in the mathematical sense meaning a one-dimensional array. We will learn in Chapter 5, Working with Matrices and ufuncs, about specialized NumPy arrays, which represent matrices. Vector a holds the squares of integers 0 to n, for instance, if n is equal to 3, then a is equal to (0,1, 4). Vector b holds the cubes of integers 0 to n, so if n is equal to 3, then b is equal to (0,1, 8). How will you do that using plain Python? After we come up with a solution, we will compare it to the NumPy equivalent.

  1. Adding vectors using pure Python: The following function solves the vector addition problem using pure Python without NumPy:

    def pythonsum(n):
       a = range(n)
       b = range(n)
       c = []
       for i in range(len(a)):
           a[i] = i ** 2
           b[i] = i ** 3
           c.append(a[i] + b[i])
       return c


    Downloading the example code files

    You can download the example code files from your account at for all the Packt Publishing books you have purchased. If you purchased this book elsewhere, you can visit and register to have the files e-mailed directly to you.

  2. Adding vectors using NumPy: Following is a function that achieves the same result with NumPy:

    def numpysum(n):
      a = np.arange(n) ** 2
      b = np.arange(n) ** 3
      c = a + b
      return c

Notice that numpysum() does not need a for loop. Also, we used the arange() function from NumPy that creates a NumPy array for us with integers 0 to n. The arange() function was imported; that is why it is prefixed with numpy (actually, it is customary to abbreviate it via an alias to np).

Now comes the fun part. The preface mentions that NumPy is faster when it comes to array operations. How much faster is NumPy, though? The following program will show us by measuring the elapsed time, in microseconds, for the numpysum() and pythonsum() functions. It also prints the last two elements of the vector sum. Let's check that we get the same answers by using Python and NumPy:


from __future__ import print_function
import sys
from datetime import datetime
import numpy as np

 Chapter 1 of NumPy Beginners Guide.
 This program demonstrates vector addition the Python way.
 Run from the command line as follows
  python n
 where n is an integer that specifies the size of the vectors.

 The first vector to be added contains the squares of 0 up to n.
 The second vector contains the cubes of 0 up to n.
 The program prints the last 2 elements of the sum and the elapsed time.

def numpysum(n):
   a = np.arange(n) ** 2
   b = np.arange(n) ** 3
   c = a + b

   return c

def pythonsum(n):
   a = range(n)
   b = range(n)
   c = []

   for i in range(len(a)):
       a[i] = i ** 2
       b[i] = i ** 3
       c.append(a[i] + b[i])

   return c

size = int(sys.argv[1])

start =
c = pythonsum(size)
delta = - start
print("The last 2 elements of the sum", c[-2:])
print("PythonSum elapsed time in microseconds", delta.microseconds)

start =
c = numpysum(size)
delta = - start
print("The last 2 elements of the sum", c[-2:])
print("NumPySum elapsed time in microseconds", delta.microseconds)

The output of the program for 1000, 2000, and 3000 vector elements is as follows:

$ python 1000
The last 2 elements of the sum [995007996, 998001000]
PythonSum elapsed time in microseconds 707
The last 2 elements of the sum [995007996 998001000]
NumPySum elapsed time in microseconds 171
$ python 2000
The last 2 elements of the sum [7980015996, 7992002000]
PythonSum elapsed time in microseconds 1420
The last 2 elements of the sum [7980015996 7992002000]
NumPySum elapsed time in microseconds 168
$ python 4000
The last 2 elements of the sum [63920031996, 63968004000]
PythonSum elapsed time in microseconds 2829
The last 2 elements of the sum [63920031996 63968004000]
NumPySum elapsed time in microseconds 274

What just happened?

Clearly, NumPy is much faster than the equivalent normal Python code. One thing is certain, we get the same results whether we use NumPy or not. However, the result printed differs in representation. Notice that the result from the numpysum() function does not have any commas. How come? Obviously, we are not dealing with a Python list but with a NumPy array. It was mentioned in the Preface that NumPy arrays are specialized data structures for numerical data. We will learn more about NumPy arrays in the next chapter.

Pop quiz – Functioning of the arange() function

Q1. What does arange(5) do?

  1. Creates a Python list of 5 elements with the values 1-5.

  2. Creates a Python list of 5 elements with the values 0-4.

  3. Creates a NumPy array with the values 1-5.

  4. Creates a NumPy array with the values 0-4.

  5. None of the above.

Have a go hero – continue the analysis

The program we used to compare the speed of NumPy and regular Python is not very scientific. We should at least repeat each measurement a couple of times. It will be nice to be able to calculate some statistics such as average times. Also, you might want to show plots of the measurements to friends and colleagues.


Hints to help can be found in the online documentation and the resources listed at the end of this chapter. NumPy has statistical functions that can calculate averages for you. I recommend using matplotlib to produce plots. Chapter 9, Plotting with matplotlib, gives a quick overview of matplotlib.


IPython – an interactive shell

Scientists and engineers are used to experiment. Scientists created IPython with experimentation in mind. Many view the interactive environment that IPython provides as a direct answer to MATLAB, Mathematica, and Maple. You can find more information, including installation instructions, at

IPython is free, open source, and available for Linux, UNIX, Mac OS X, and Windows. The IPython authors only request that you cite IPython in any scientific work that uses IPython. The following is a list of the basic IPython features:

  • Tab completion

  • History mechanism

  • Inline editing

  • Ability to call external Python scripts with %run

  • Access to system commands

  • Pylab switch

  • Access to Python debugger and profiler

The Pylab switch imports all the SciPy, NumPy, and matplotlib packages. Without this switch, we will have to import every package we need ourselves.

All we need to do is enter the following instruction on the command line:

$ ipython --pylab
IPython 2.4.1 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.
Using matplotlib backend: MacOSX

In [1]: quit()

The quit()command or Ctrl + D quits the IPython shell. We may want to be able to go back to our experiments. In IPython, it is easy to save a session for later:

In [1]: %logstart
Activating auto-logging. Current session state plus future input saved.
Filename       :
Mode           : rotate
Output logging : False
Raw input log  : False
Timestamping   : False
State          : active

Let's say we have the vector addition program that we made in the current directory. Run the script as follows:

In [1]: ls
In [2]: %run -i 1000

As you probably remember, 1000 specifies the number of elements in a vector. The -d switch of %run starts an ipdb debugger with c the script is started. n steps through the code. Typing quit at the ipdb prompt exits the debugger:

In [2]: %run -d 1000
*** Blank or comment
*** Blank or comment
Breakpoint 1 at: /Users/…/


Enter c at the ipdb> prompt to start your script.

ipdb> c
> /Users/…/<module>()
1---> 3 import sys
      4 from datetime import datetime
ipdb> n
1     3 import sys
----> 4 from datetime import datetime
      5 import numpy
ipdb> n
> /Users/…/<module>()
      4 from datetime import datetime
----> 5 import numpy
ipdb> quit

We can also profile our script by passing the -p option to %run:

In [4]: %run -p 1000
         1058 function calls (1054 primitive calls) in 0.002 CPU seconds
   Ordered by: internal time
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.001    0.001   0.001    0.001
1 0.001    0.001   0.002    0.002 {execfile}
1000 0.000    0.0000.0000.000 {method 'append' of 'list' objects}
1 0.000    0.000    0.002    0.002<module>)
1 0.000    0.0000.0000.000
3    0.000    0.0000.0000.000 {range}
1    0.000    0.0000.0000.000
3/1    0.000    0.0000.0000.000
2    0.000    0.0000.0000.000 {method 'reduce' of 'numpy.ufunc' objects}
4    0.000    0.0000.0000.000 {built-in method now}
2    0.000    0.0000.0000.000
2    0.000    0.0000.0000.000 {numpy.core.multiarray.arange}
1    0.000    0.0000.0000.000
3/1    0.000    0.0000.0000.000
1    0.000    0.0000.0000.000
2    0.000    0.0000.0000.000
1    0.000    0.0000.0000.000
2    0.000    0.0000.0000.000<lambda>)
1    0.000    0.000    0.002    0.002<string>:1(<module>)
11    0.000    0.0000.0000.000 {len}
2    0.000    0.0000.0000.000 {isinstance}
1    0.000    0.0000.0000.000 {reduce}
1    0.000    0.0000.0000.000 {method 'ravel' of 'numpy.ndarray' objects}
4    0.000    0.0000.0000.000 {method 'rstrip' of 'str' objects}
3    0.000    0.0000.0000.000 {issubclass}
2    0.000    0.0000.0000.000 {method 'item' of 'numpy.ndarray' objects}
1    0.000    0.0000.0000.000 {max}
1    0.000    0.0000.0000.000 {method 'disable' of '_lsprof.Profiler' objects}

This gives us a bit more insight in to the workings of our program. In addition, we can now identify performance bottlenecks. The %hist command shows the commands history:

In [2]: a=2+2
In [3]: a
Out[3]: 4
In [4]: %hist
1: _ip.magic("hist ")
2: a=2+2
3: a

I hope you agree that IPython is a really useful tool!


Online resources and help

When we are in IPython's pylab mode, we can open manual pages for NumPy functions with the help command. It is not necessary to know the name of a function. We can type a few characters and then let tab completion do its work. Let's, for instance, browse the available information for the arange() function:

In [2]: help ar<Tab>
In [2]: help arange

Another option is to put a question mark behind the function name:

In [3]: arange?

The main documentation website for NumPy and SciPy is at Through this web page, we can browse the NumPy reference at, the user guide, and several tutorials.

The popular Stack Overflow software development forum has hundreds of questions tagged numpy. To view them, go to

If you are really stuck with a problem or you want to be kept informed of NumPy development, you can subscribe to the NumPy discussion mailing list. The e-mail address is . The number of e-mails per day is not too high with almost no spam to speak of. Most importantly, the developers actively involved with NumPy also answer questions asked on the discussion group. The complete list can be found at

For IRC users, there is an IRC channel on irc:// The channel is called #scipy, but you can also ask NumPy questions since SciPy users also have knowledge of NumPy, as SciPy is based on NumPy. There are at least 50 members on the SciPy channel at all times.



In this chapter, we installed NumPy and other recommended software that we will be using in some sections of this book. We got a vector addition program working and convinced ourselves that NumPy has superior performance. You were introduced to the IPython interactive shell. In addition, you explored the available NumPy documentation and online resources.

In the next chapter, you will take a look under the hood and explore some fundamental concepts including arrays and data types.

About the Author
  • Ivan Idris

    Ivan Idris has an MSc in experimental physics. His graduation thesis had a strong emphasis on applied computer science. After graduating, he worked for several companies as a Java developer, data warehouse developer, and QA analyst. His main professional interests are business intelligence, big data, and cloud computing. Ivan Idris enjoys writing clean, testable code and interesting technical articles. Ivan Idris is the author of NumPy 1.5. Beginner's Guide and NumPy Cookbook by Packt Publishing.

    Browse publications by this author
Latest Reviews (3 reviews total)
It is good book for beginners in Data cirnce.
This book is very informative, and well written. After completing just the first few chapters I am already programming in NumPy!
NumPy: Beginner's Guide - Third Edition
Unlock this book and the full library FREE for 7 days
Start now