Introduction to Machine Learning

Another powerful data structure is a dictionary. Python implements a dictionary as a built-in dict type. You may have encountered these in the C++ standard library as maps. These structures also go by other names in different languages: associative arrays, hashtables, etc.

In case you are not familiar, a dictionary acts as a lookup table.

It is made up of (key, value) pairs. You use a unique key to retrieve its corresponding value.

An example use case for a dictionary would be to retrieve a student’s name given the student’s ID.

Here is an example of how you would define a dict:

>>> student_dict = {"00-01-30": "Ali", "00-02-11": "Simon", 
...                 "00-05-67": "Francesca", "00-09-88": "Cho"
...                }
>>> print(student_dict)

Alternatively:

>>> student_dict = {}                   # or student_dict = dict()
>>> student_dict["00-01-30"] = "Ali"
>>> student_dict["00-02-11"] = "Simon"
>>> student_dict["00-05-67"] = "Francesca"
>>> student_dict["00-09-88"] = "Cho"
>>> print(student_dict)

Question: What happens if you subsequently add student_dict["00-01-30"] = "Josiah"? Will Python allow this? What will happen to "Ali"? Test it out to confirm!

Retrieving an item from a dictionary

As you would expect, you can retrieve the value of a dictionary given a key.

>>> best_student = student_dict["00-02-11"]
>>> print(best_student)

But what if the key cannot be found? What happens?

>>> student_dict["00-11-22"]    # What error message do you get? 

So you should check whether the key exists before trying to use the key to retrieve a value.

>>> if "00-11-22" in student_dict:
...    active_student = student_dict["00-11-22"]

Alternatively, you can use dict’s get() method to retrieve a value. The method will conveniently assign a default value if the key does not exist.

>>> real_student = student_dict.get("00-09-88", "John Doe")
>>> print(real_student)
>>> imaginary_student = student_dict.get("11-22-33", "John Doe")
>>> print(imaginary_student)

Other useful methods

If you need the complete list of keys:

>>> ids = student_dict.keys()
>>> print(ids)
>>> print(type(ids))
>>> ids[0]              # Can you do this?

If you need to access the elements in the dict_keys instance, you can cast it into a list. In fact, you can just cast the dictionary directly into a list to achieve the same thing.

>>> id_list = list(ids)
>>> print(id_list[0])
>>>
>>> id_list = list(student_dict)
>>> print(id_list[0])

Similarly, dict’s values() method give you the list of values.

There is also the items() method which give you a list of (key, value) pairs.

Test them out!

>>> print(student_dict.values())
>>> print(student_dict.items())