Python Programming for Biology: Bioinformatics and Beyond

Error exceptions

Python deals with errors caused by illegal circumstances, which are called exceptions, in a

relatively  graceful  manner.  Note  that  it  is  commonplace  to  say  ‘throw  an  exception’  to

mean that an exception was generated by some illegal program state. Here is a very simple

example of something that will generate an exception object, indicating the type of error,

in this case because it is illegal to divide a number by zero:

x = 1/0 # Throws a ZeroDivisionError

Catching exceptions

Rather  than  simply  letting  such  things  always  break  the  program,  Python  allows  you  to

‘try’ something and then ‘catch’ any exceptions that might occur and do something about

it, and potentially let the program carry on afterwards. The simplest version is:

try:

# code block that might throw an exception

except:

# exception handling code

# maybe let the program continue after

So if an exception occurs, the program immediately stops executing the code in the try

block and instead executes the code in the except block. In the above case, every kind of

exception error is trapped, so there is no way of knowing what kind of exception it was.

However,  you  can  specify  what  kind  of  exception  is  trapped.  Python  has  many  inbuilt

exception types (classes). As illustrated above, if you try and divide by 0 then you get a

ZeroDivisionError, and you can check specifically for that:

x = 1

y = 0

---

try:

z = x + y

w = x / y

t = x * y

except ZeroDivisionError:

print('divided by zero')

print('program did not stop')

In this specific example, an exception occurs in the calculation of w, so it is not set and,

further, the calculation of t never happens and an error message is printed instead. Because

the illegal division by zero was dealt with the program can then continue. You can even

get hold of an exception object, which can be printed out to show the internal Python error

message:

x = 1

y = 0

try:

w = x / y

except ZeroDivisionError as errorObj:

print(errorObj) # prints 'integer division or modulo by zero'

print('program did not stop')

Before Python 2.6 the syntax for catching the exception object was different, so if you

are  using  an  older  version  of  Python  then  you  have  to  use  a  comma  instead  of  ‘as’  (the

comma syntax also works in Python 2.6 and 2.7 but not in Python 3):

try:

w = x / y

except ZeroDivisionError, errorObj:

print(errorObj) # prints 'integer division or modulo by zero'

This  reports  the  Python  message  about  what  happened,  which  you  would  get  if  the

program  failed,  but  naturally  because  we  used  the  try:  clause  to  catch  the  problem  the

program did not stop. There is a simple ‘base’ exception in Python called Exception, and

that lets you trap all types of exception and get hold of the exception object:

try:

# code block that might throw an exception

except Exception as errorObj:

# exception handling code

Also, you can catch different types of exception by repeating the except statement. This

allows  different  problems  to  be  handled  in  different  ways.  For  example,  if  you  try  to

divide  something  by  a  text  string  (or  a  string  by  something)  in  Python  you  get  a

TypeError.  So  your  division  code  could  check  for  this  and  for  zero  division,  printing  an

appropriate message for each situation:

try:

w = x / y

except ZeroDivisionError as errorObj:

print('divided by zero', errorObj)

except TypeError as errorObj:

---

print('divided by something silly', errorObj)