Appendix – What for?#
Contributed by Audra McNamee
Python offers convenient and concise ways to write loops, but
similarity between different patterns of for
loops often
confuses students. Using one pattern when another is
required often leads to TypeError
exceptions.
This brief walkthrough will help you distinguish between
for i in range(len(array)):
and
for i in array:
A list of lists#
Throughout this walkthrough,
assume array = [[ item1, item2], [item3, item4], [item5, item6 ]]
.
We visualize this list of lists as being arranged in rows and columns:
array = [[item1, item2],
[item3, item4],
[item5, item6]]
We say array
has three rows and two columns.
Iteration through the items in a list#
In Python, lists (and strings) are iterable objects. Thus, it’s perfectly valid to write
for row in array:
In first iteration, row
is [item1, item2]
.
In the second iteration, row
is [item3, item4]
.
In the third iteration, row
is [item5, item6]
.
Note that when we iterate through elements of a list, we get the elements themselves, and not their positions.
Iteration through the indices of a list#
When you require an item’s position as well as its value, iteration with
for row in range(len(array)):
is appropriate.
In the first iteration, row
is 0
.
In the second iteration, row
is 1
.
In the third iteration, row
is 2
.
TypeError from mistaken identity#
Type errors typically result from iterating through loop elements when we need their indices, or using indices as if they were elements.
Say you intend to loop through an object by column, and write
for col in array[0]:
for row in array:
print(array[row][col])
The third line, print(array[row][col])
,
will produce an IndexError after the first iteration.
col
is item1
and row
is [item1, item2, item3]
.
To access array[row][col]
we would need row
and col
to be indices, but instead col
is an element of
the list.
array[[item1, item2, item3]][item1]
is a type error.
Since we needed indices for array[row][col]
, in
this case we should use for
loops that iterate through
list indices:
for col in range(len(array[0])):
for row in range(len(array)):
print(array[row][col])
In the first iteration col
is 0
and row
is 0
.
array[0][0]
is the element in row 0
and and column 0
,
so item1
will be printed.
The takeaway: always know whether you’re iterating by item or by index.
IndexError from using the wrong loop bound#
Say you attempt to iterate through array
by columns with this code:
for col in range(len(array)):
for row in range(len(array)):
print(array[row][col])
This code will print a few elements correctly, then crash. Try to spot the error before reading on.
Which bound is which#
The code above prints item1
, …, item6
then crashes with IndexError
.
This is because range(len(array))
iterates from 0 to 2, but there is no column 2.
When the loop attempts
print(array[0][2])
Python raisesIndexError
.
When we write loops to iterate through the indices
of list, we need to make sure the bounds of the
loop match the list we are iterating through.
In the case above, col
ranges through
range(len(array))
, the indexes of the rows,
when we need the indexes of the columns.
Correct code to print each element could be
for col in range(len(array[0])):
for row in range(len(array)):
print(array[row][col])
Now len(array[0])
is the number of columns,
and len(array)
is the number of rows, so
array[row][col]
will always be a valid combination
of row index and column index (provided each row
is the same length).
Handle empty lists#
If the length of array
could equal 0, you must be
careful not to access the first column (column 0) if
there is no such column. The code above starting with
for col in range(len(array[0])):
will crash if array
is [ ]
. For this case
we need a guard:
if len(array) == 0:
# Do the appropriate thing for the empty array
else:
for col in range(len(array[0])):
...
Often the appropriate action will be to return a default value, and we can simplify this to
if len(array) == 0:
return appropriate_value
for col in range(len(array[0])):
...
Mnemonic loop variable names#
A mnemonic is a memory aid. One of the most valuable mnemonic aids available to programmers is variable names. Naming variables well can help us keep track of how we are controlling loops and prevent inconsistency.
If we loop through the elements of a list, we can use variable names that indicate that as clearly as possible. If the rows of our array were teams, and the columns were members of each team, then our loops might have been written
for team in array:
for member in team:
If we loop through the indexes of the elements, it can help to use variable names that remind us of that:
for team_i in range(len(array)):
for member_i in range(len(array[0])):
Sometimes introducing variables for both the index and the element is useful in keeping the distinction clear:
# Iteration by row
for row_i in range(len(array)):
row = array[row_i]
for col_i in range(len(row)):
col = row[col_i]
...
or
# Iteration by column
if len(array) == 0:
# Return the appropriate value
for col_i in range(len(array[0])):
for row_i in range(len(array)):
row = array[row_i]
item = row[col_i]
...
This code is slightly more verbose than code without the extra variables, but the extra lines are worthwhile if they make the code clearer.
Poisonous variable names#
At the very least we should avoid using misleading variable names. Never ever ever do this:
for row in range(len(array[0])):
for col in range(len(array)):
print(array[col][row])
# BURN THIS CODE. BURY IT. WEAR GLOVES.
Summary#
Python gives you convenient choices in patterns
for for
loops, but you need
to choose carefully and consistently.
Use
for element in array:
to loop through list elements, orfor element in range(len(array))
to loop through indices.Use the right list to determine the range of indices.
If the outer loop iterates through indices of an inner list, you will probably need to write a special case handler for an empty outer list.
Use variable names that help you remember which list is which, and whether you are iterating through list elements or list indices.
When we write
a[x][y]
,x
is always the index of the outer list andy
is always the index of the inner list, no matter what they are called.