How to catch conditional exceptions in python? (try, except, pass)

Python is a highly structured language and it is generally considered unnecessary and bad practice in these languages ​​to use the unconditional transfer statement. In this case, Python does not directly have instructions like gotoor jump.

These problems are solved in a simple way using structured programming features (conditionals, loops, etc). In this case it is solved simply by placing another returnin the try, for example:

def get_part(string):     
    try:
        x = string.split("-")
        return x[1].split("[")[0]

    except IndexError:
        return string.split("[")[1][:-1]

You can have as many returnas you want in a function and at any point (inside loops, inside conditionals, etc). The moment one is executed the function ends at that point. Although it is not usually the most indicated, in Python you can have different returnones that return different types even.

Another option is to not use exception handling and use conditionals:

def GetPart(string): 
    x = string.split("-") 

    if len(x) != 2:        # Si x no tiene dos elementos (cadena sin un "-")
        y = string.split("[")
        if len(y) == 2:
            return y[1][:-1]

    y = x[1].split("[")

    if 1 <= len(y) <= 2:
        return y[0]

    return None           # Para cualquier otra cadena

>>> get_part("Parte1[Parte2]")
'Parte2'
>>> get_part("Parte1-Parte2[Part3]")
'Parte2'

Edit clarifying comment question:

You can nest blocks try-exceptwithout problems, however, since you are inside a function and any function returnterminates its execution, you can not nest them as long as each one has its returnin the try, for example:

def to_float(x):
    try:
        f = float(x)
        return f
    except:
        pass

    # Cadena con coma como separador
    try:
        f = float(x.replace(",", "."))
        return f
    except:
        pass

    # Cadena con fraccíon
    try:
        dividendo, divisor = (int(n) for n in x.split("/"))
        return dividendo / divisor
    except:
        pass

    raise ValueError("could not convert {} to float".format(repr(x)))

>>> to_float(13)
13.0
>>> to_float("13,5")
13.5
>>> to_float("13.5")
13.5

>>> to_float("13-5")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "test.py", line 16, in to_float
    raise ValueError("could not convert {} to float".format(repr(x)))
ValueError: could not convert '13-5' to float

>>> to_float("2/3")    
0.6666666666666666    

EXHIBIT:

Catching exceptions in Python is done using the statement try, which basically consists of:

• Clausetry : includes the code whose possible exceptions we are going to handle.

• Clause or clausesexcept : that contain the code to be executed in case of tryfailure. We can specify that a exceptsingle catch certain types of exceptions:

  try:
      ...
  except ValueError:
      ...
  
  ################################   
  
  try:
      ...
  except (ValueError, MathError):
  ...  
  
  ################################    
  
  try:
      ...
  except ValueError:
      ...
  except MathError:
      ...
  

  To catch an exception in the exceptuse as:

  try:
      ...
  except Exception as e:
      print(e)
  
  ################################
  
  try:
      ...
  except (ValueError, MathError) as e:
      print(e)
  

• Optional clauseelse : is executed when the trycompletes with no exceptions. In this case the contained code is not protected by tryavoiding catching exceptions that we don't want to handle.

• Optional clause finally : always executed before terminal block, whether or not an exception occurred. When an exception has occurred in the clause tryand has not been handled by a clause exceptit is re-raised after the clause has been executed finally. Exit statements such as break, continueor returndo not prevent finallyexecution before executing them .

The statement passjust "does nothing", is used when we declare a block that cannot be syntactically empty (causes a IdentationError) like if, elif, else, except, defor class, but we don't want it to do anything, at least for now.