From lists to arrays

What you will be able to do

After this lesson you can

• explain why scientists use NumPy arrays and not lists,
• make arrays with np.array, np.arange, np.linspace, np.zeros, np.ones and np.full,
• read the dtype, the shape and the length of an array, and
• change the type of an array.

Why arrays?

A science experiment gives numbers, and often many of them. A sensor writes a temperature every second, and in one hour there are 3600 numbers. A telescope records millions. You want to find the average, to convert the units, to find the biggest, and to draw a graph. With a list, you must write a loop for every one of these jobs.

NumPy (the name comes from Numerical Python) is the library that solves this. Its main object is the array: a row of numbers, all of the same type, on which you can do maths all at once, with no loop. Almost all scientific work in Python, from physics to biology, is built on NumPy. The libraries that you meet later in this book, SciPy, matplotlib and pandas, all use its arrays.

On this site, NumPy is ready to use. The first time that your code uses it, the site needs a few seconds to load it. After that, it is quick. By custom, we write the import in this way:

np is a short name, and every NumPy program in the world uses it.

A list and an array

Predict first. Look at the four lines that print. Which two print the same thing? Then run it.

An array does what a mathematician expects. This is the whole idea.

Making arrays

Which one to choose, arange or linspace? A good rule: use arange for whole-number steps (0, 2, 4, ...), and linspace when you need exactly a number of points, or when the step is a decimal. With a decimal step, arange can give one number too many or too few, because decimals are not exact. linspace always gives the number that you asked for.

Use linspace when you know how many points you want, as for a smooth graph. Use arange when you know the step.

One type for all the numbers

The one type is what makes arrays fast. NumPy knows that every number takes the same space, so it can work through them quickly.

How much faster?

Common mistakes

• Forgetting import numpy as np. Then np is not defined, and Python says NameError.
• Giving np.array the numbers one by one. It needs one list: np.array([1, 2, 3]).
• Thinking that arange includes the stop. np.arange(0, 10) ends at 9.
• *Using ` on a list and expecting maths.** For a list, *` repeats. Make an array first.
• Mixing types and being surprised. One decimal turns the whole array into decimals.
• Thinking that astype(int) rounds. It cuts off the decimal part.

Practice

Start with two quick questions. Each New numbers gives you a fresh one, so you can practise as long as you like.

Then the labs and the problems:

1. Warm-up. Open the lab Make the arrays. You make four arrays, and the checker tests them.
2. Build. Open the problem List or array? and the problem Build the arrays.

Remember these

• A NumPy array holds many numbers of one type, and does maths on all of them at once.
• np.array(list), np.arange(start, stop, step), np.linspace(start, stop, count), np.zeros, np.ones, np.full.
• dtype, shape, ndim and len describe an array.
• astype changes the type. One decimal makes the whole array decimals.

Check yourself

Try each one on your own first. If you are stuck, open a hint. Then check with Show the answer.

1. What does np.array([1, 2]) + np.array([3, 4]) give? And [1, 2] + [3, 4]?

2. How many numbers does np.arange(0, 1, 0.25) give, and does it include 1?

3. How many numbers does np.linspace(0, 1, 11) give, and what is the step?

4. What is the type of np.array([1, 2.0, 3])?