Reading error messages and a small project

A Python traceback is read from the bottom: the last line names the error, and the line above the code shows where it happened.

A moth taped into a log book in 1947

Today most bugs are mistakes in the code, and Python is good at telling you about the ones it can detect. Beginners often see a red error as a verdict. It is closer to a delivery note: what went wrong, on which line, and in which piece of code. The trick is knowing where to look.

Read a traceback from the bottom up

total = 0
count = 0
print(total / count)

Traceback (most recent call last):
  File "main.py", line 3, in <module>
    print(total / count)
          ~~~~~~^~~~~~~
ZeroDivisionError: division by zero

Start with the last line. ZeroDivisionError: division by zero gives the type of error and a short explanation. Then move up to the File line: the file name (the full folder path appears there on your computer; it is shortened to main.py here) and the line number, 3. The line under that is the code itself, and Python 3.12 marks the exact operation that failed with ^ under the / sign. Only after reading all three should you start changing things. When the error happens inside a function, the traceback lists every call on the way there, and the bottom-most File line is usually the one to look at first.

A SyntaxError is different from the rest. Python finds it before running anything, so not even line 1 runs, and there is no "Traceback" header. All the others happen while the program runs, after the earlier lines have already done their work. When you expect a particular error, you can catch it with try and except, which the tutorial covers in its chapter on errors: Python tries the indented block, and if that error happens, it runs the except block instead of stopping.

print("before")
n = int("abc")
print("after")

before
Traceback (most recent call last):
  File "main.py", line 2, in <module>
    n = int("abc")
        ^^^^^^^^^^
ValueError: invalid literal for int() with base 10: 'abc'

text = "abc"
try:
    n = int(text)
except ValueError:
    n = 0
print(n)

0

The hardest bugs give no message at all. A logic error is a program that runs and prints the wrong answer, like the grading program from the unit on if that gave 85 a C. Python cannot know what you meant. For these, go back to the trace table, or add temporary print lines that show the variables at each step, then remove them once the bug is found.

Small project: a pocket-money log

The project uses something from nearly every unit: a dictionary of spending, a function that returns a total, a for loop, an f-string and an if. Type it in, run it, then make it yours: change the items, add a category, or warn when any single item costs more than half the allowance.

def total(spent):
    t = 0
    for amount in spent.values():
        t += amount
    return t

allowance = 30000
spent = {"snacks": 4500, "bus card": 10000, "notebook": 2500}
used = total(spent)
left = allowance - used
for item, amount in spent.items():
    print(f"{item}: {amount} won")
print(f"Spent {used}, left {left}")
if left < 15000:
    print("Less than half left")

snacks: 4500 won
bus card: 10000 won
notebook: 2500 won
Spent 17000, left 13000
Less than half left

Before you change it, write a trace table for used and left, and predict what happens if you add a 6,000-won item. Then debug on purpose: delete the colon after the def line, misspell spent once, and divide used by zero. Read each traceback from the bottom before you fix it. Being able to break a program and mend it again is the real skill this course has been building.