Comprehensible Comprehensions

Assumptions:

• Know how for loops work.
• You are not a fan of list comprehensions.

iterable:

• anything you can iterate over.

generator:

• lazy iterable.

List comprehensions

• Turn one list into another list.
• Making new list into old list more common in Python.
• Often copy old list to new list to avoid mutating old list.
• Contains some information as for loop.
• Are more specific about intent of code then for loop.

Complaint

• Makes readable for loops unreadable.

  • Comprehensions often written on one line. Very condensed. No white space.

  • Operation, Loop, Condition. Don’t have to be on one line:

    .. code-block:: python

    squared_odds = [ n ** 2 for n in numbers if n %2 == 1 ]

  • Should always start by writing comprehension on multiple lines.

Generator expressions

• If you are only going to loop over result once, you should use a generator expression instead of a list comprehension.

  .. code-block:: python

  squared_numbers = (n ** 2 for n in numbers)

• Generators are lazy, only generate on demand.
• Generators are one time only. Can only loop over once. Single use.

When not to use comprehensions

• Only for turning one list into another list.
• Do not have side effects, e.g. print, in comprehension.
• When dict/list constructors can be used instead.
• When not changing or filtering anything.