Dictionary based switch-like statement with actions

Question

Hi,

I'm relatively new to Python and would like to know if I'm reinventing a wheel or do things in a non-pythonic way - read wrong.

I'm rewriting some parser originally written in Lua. There is one function which accepts a field name from imported table and its value, does some actions on value and stores it in target dictionary under a appropriate key name.

In the original code it's solved by long switch-like statement with anonymous functions as actions. Python code looks like the following:

class TransformTable:
    target_dict = {}
    ...
    def mapfield(self, fieldname, value):
        try:
            {
                'productid': self.fn_prodid,
                'name': self.fn_name,
                'description': self.fn_desc,
                ...
            }[fieldname](value)
        except KeyError:
            sys.stderr.write('Unknown key !\n')

    def fn_name(val):
        validity_check(val)
        target_dict['Product'] = val.strip().capitalize()
    ...

Every "field-handler" function does different actions and stores in different keys in target_dict, of course. Because Python does not support anonymous functions with statements (or did I missed something ?) the functions have to be written separately which does code less readable and unnecessarily complicated.

Any hints how to do such tasks in a more elegant and more pythonic way are appreciated.

Thx

David

Answer 1

If by any means possible, you could name your member functions based on the field names and just do something like this:

getattr(self, "fn_" + fieldname)(value)

Edit: And you can use hasattr to check if the function exists, instead of expecting a KeyError. Or expect an AttributeError. At any rate, you should put only the access inside your try..except, and call it outside, since otherwise a KeyError caused within one of the field methods could get misunderstood.

Answer 2

I applied an approach similar to @Matti Virkkunen'a a while ago in my answer to a question titled "switch case in python doesn't work; need another pattern". It also demonstrates a relatively easy and graceful way of handling unknown fields. Transliterated to terms in your example it would look like this:

class TransformTable:
    target_dict = {}

    def productid(self, value):
        ...
    def name(self, value):
        validity_check(value)
        self.target_dict['Product'] = value.strip().capitalize()
    def description(self, value):
        ...

    def _default(self, value):
        sys.stderr.write('Unknown key!\n')

    def __call__(self, fieldname, value):
        getattr(self, fieldname, self._default)(value)

transformtable = TransformTable() # create callable instance

transformtable(fieldname, value) # use it