Scientific Python antipatterns advent calendar day seven

For today, a simple error that it easy to fall into. As a reminder, I’ll post one tiny example per day with the intention that they should only take a couple of minutes to read.

If you want to read them all but can’t be bothered checking this website each day, sign up for the mailing list:

and I’ll send a single email at the end with links to them all.

Iterating over dictionaries without using `items`

In Python we use dictionaries when we want to be able to efficiently look up the value associated with a key:

# create a dictionary
emails = {
    'Martin' : 'martin@pythonforbiologists.com', # only this one is real!
    'Jess'   : 'jessica.rodriguez@gmail.com',
    'Phil'   : 'p.smith@outlook.com'
}

# look up an email
emails['Jess']

'jessica.rodriguez@gmail.com'

Sometimes we need to do something to all the key/value pairs in a dictionary. If we’re not sure how to do this, the obvious thing to try is iterating over the dictionary:

for x in emails:
    print(x)

Martin
Jess
Phil

From the output, we will quickly realise that iterating over a dictionary gives us the keys, so we can add a step to get the matching values:

for name in emails:
    email_address = emails[name]
    print(name, email_address, sep='\t')

Martin  martin@pythonforbiologists.com
Jess    jessica.rodriguez@gmail.com
Phil    p.smith@outlook.com

This works, but is harder than it needs to be. Whenever you see this pattern - a loop that iterates over keys, then inside the loop a line that gets the matching value - we can replace it with a call to items:

emails.items()

dict_items([('Martin', 'martin@pythonforbiologists.com'), ('Jess', 'jessica.rodriguez@gmail.com'), ('Phil', 'p.smith@outlook.com')])

The items method gives us a list where each elemnt is a tuple containing the key and the value. So we can set up the loop in a single line:

for name, email_address in emails.items():
    print(name, email_address, sep='\t')

Martin  martin@pythonforbiologists.com
Jess    jessica.rodriguez@gmail.com
Phil    p.smith@outlook.com

which normally makes the body of the loop clearer. Occasionally you might see this:

for item in emails.items():
    name, email_address = item
    print(name, email_address, sep='\t')

Martin  martin@pythonforbiologists.com
Jess    jessica.rodriguez@gmail.com
Phil    p.smith@outlook.com

or even this:

for item in emails.items():
    name          = item[0]
    email_address = item[1]
    print(name, email_address, sep='\t')

Martin  martin@pythonforbiologists.com
Jess    jessica.rodriguez@gmail.com
Phil    p.smith@outlook.com

but both are more clumly than doing the unpacking in a loop.

Bonus: it’s only necessary to use a dictionary if we want to be able to loop up individual keys. If we find ourselves building a dictionary, but only ever using it in a loop, then it should probably be a list of tuples instead:

emails = [
    ('Martin','martin@pythonforbiologists.com'), # only this one is real!
    ('Jess','jessica.rodriguez@gmail.com'),
    ('Phil','p.smith@outlook.com')
]

which will be even easier to iterate over:

for name, email_address in emails:
    print(name, email_address, sep='\t')

Martin  martin@pythonforbiologists.com
Jess    jessica.rodriguez@gmail.com
Phil    p.smith@outlook.com

and much more memory efficient.

One more time; if you want to see the rest of these little write-ups, sign up for the mailing list:

Iterating over dictionaries without using items

Iterating over dictionaries without using `items`