for loops

Sources

This lesson is based on the Software Carpentry group’s lessons on Programming with Python.

Basics of for loops

  1. Loops allow parts of code to be repeated over some number of times. One of the simple loop options in Python is the for loop.

    >>> word = 'rock'
>>> for char in word:
...    print(char)
...
r
o
c
k

  2. for loops in Python have the general form below.

    for variable in collection:
    do things with variable

    The variable can be any name you like, and the statement of the for loop must end with a :. The code that should be executed as part of the loop must be indented beneath the for loop, and the typical indentation is 4 spaces. There is not additional special word needed to end the loop, just change the indentation back to normal.

  3. Let’s consider another example.

    >>> length = 0
>>> for letter in 'earthquake':
...    length = length + 1
...
>>> print('There are', length, 'letters')
There are 10 letters

    Note that the variable used in the loop, letter is just a normal variable and still exists after the loop has completed with the final value given to letter.

    >>> print('After the loop, letter is', letter)
e

  4. A loop can be used to iterate over any list of values in Python. So far we have considered only character strings, but we could also write a loop that performs a calculation a specified number of times.

    >>> for number in range(5):
...     print(number)
...
0
1
2
3
4

    What happens here? Well, in this case, we use a special function called range() to give us a list of 5 numbers [0, 1, 2, 3, 4] and then print each number in the list to the screen. When given an integer (whole number) as an argument, range() will produce a list of numbers with a length equal to the specified number. The list starts at zero and ends with number-1.

  5. Often when you use for loops, you are looping over the values in a list and either calculating a new value or modifying the existing values. Let’s consider an example.

    >>> mylist = [0.0, 1.0, 2.0, 3.0, 4.0, 5.0]
>>> print(mylist)
[0.0, 1.0, 2.0, 3.0, 4.0, 5.0]
>>> for i in range(6):
...     mylist[i] = mylist[i] + i
...
>>> print(mylist)
[0.0, 2.0, 4.0, 6.0, 8.0, 10.0]

    So, what happened? We first create a list of 6 numbers. Then, we loop over 6 values using the range() function and add each value to the existing location in mylist.

  6. One of the drawbacks in the example above is that we need to know the length of the list before running that for loop example. However, we already know how to find the length of a list using the len() function, and we can take advantage of this knowledge to make our for loop more flexible.

    >>> for i in range(len(mylist)):
...     mylist[i] = mylist[i] + i
...
>>> print(mylist)
[0.0, 3.0, 6.0, 9.0, 12.0, 15.0]

    Using the len() function with range() to perform calcluations using list or array values is an extremely common operation in Python.

Exercise - Putting it together

  • Create a new NumPy array called numbers that starts at 1 and goes to 100 in increments of 1
  • Create a new NumPy array of zeros called squared that is the same size as numbers
  • Using a for loop, calculate the square of each value in numbers and store it in the corresponding location in squared

Exercise - Let’s get functional

  • Take your code above and use it to create a new Python function square() that accepts a NumPy array and returns an array of squared values
  • Do you get the expected results when using your function?
  • Can you break your function (get it to give an error message)? If so, how?

Exercise - Drag race

IPython has a magic function called %timeit that you can use to calculate how long it takes a line of code (or program) to execute.

>>> %timeit np.ones(100000000).mean()
loop, best of 3: 427 ms per loop

We can use this now to compare the performance of your new square() function with calculating the square of values directly in NumPy

  • Create a new NumPy array called input that goes from 1 to 10 in increments of 0.0000001
  • Use %timeit with your function above to calculate the square of input, storing the output in an array called out1
  • Compare the performance of your function to simply squaring the input array directly and storing its output as out2
  • Can you see any benefits to using NumPy?