How can I dynamically generate class instances with single attributes read from flat file in Python?

Question

I apologise if this question has already been asked.

I'm really new to Python programming, and what I need to do is this: I have a .csv file in which each line represent a person and each column represents a variable. This .csv file comes from an agent-based C++ simulation I have done. Now, I need to read each line of this file and for each line generate a new instance of the class Person(), passing as arguments every variable line by line.

My problem is this: what is the most pythonic way of generating these agents while keeping their unique ID (which is one of the attributes I want to read from the file)? Do you suggest creating a class dictionary for accessing every instance? But I still need to provide a name to every single instance, right? How can I do that dynamically? Probably the best thing would be to use the unique ID, read from the file, as the instance name, but I think that numbers can't be used as instance names, can they? I miss pointers! :(

I am sure there is a pythonic solution I cannot see, as I still have to rewire my mind a bit to think in pythonic ways... Thank you very much, any help would be greatly appreciated! And please remember that this is my first project in python, so go easy on me! ;)

EDIT: Thank you very much for your answers, but I still haven't got an answer on the main point: how to create an instance of my class Person() for every line in my csv file. I would like to do that automatically! Is it possible? Why do I need this? Because I need to create networks of these people with networkx and I would like to have "agents" linked in a network structure, not just dictionary items.

score 3 · Accepted Answer · edited May 23 '17 at 12:34

3

For reading the csv file and generate a dict file from it you should definitively have a look at http://docs.python.org/library/csv.html#csv.DictReader.

If you have a csv file that has as a first row the name of the fields and the data on the rest of the rows, DictReader would generate a dictionary taking as keys the name of the fields defined in the first row. To give an example if you give it the following csv file:

Field1,Field2,Field3
1,2,3
4,5,6

It would return the following list:

[{'Field1':1,'Field2':2,'Field3':3} , {'Field1':4,'Field2':5,'Field3':6} ]

ADDED:

Regarding the creation of the instances based on the dictionary you should probably have a look at the following:

Creating class instance properties from a dictionary in Python

Or taken from the following link you can do the following:

class Person:
    def __init__(self, **kwds):
        self.__dict__.update(kwds)

# that's it!  Now, you can create a Bunch
# whenever you want to group a few variables:

person = Person(field1=1, field2=2, field3=3)

Where you could use every entry of the dictionary returned by the dictReader to create a new instance:

reader = csv.DictReader(open('yourfile.csv', 'rU'), delimiter=',')
entries = [Person(**entry) for entry in reader]

edited May 23 '17 at 12:34

Community

1
1

answered May 14 '12 at 12:26

txominpelu

1,067
1
6
11

Thank you. Ny the way do you know any way of creating an instance for each line of my csv file automatically? – Wilco May 15 '12 at 16:33
I think the latest update may give you a complete explanation. Tell me if you have any doubt. – txominpelu May 15 '12 at 17:30
In fact, I need however a way to directly access an instance by key (a unique ID already contained in my csv file). Probably it's better if I create a new post for this? However, I need to create a network of people with networkx. The simplest thing I want to do is to link agents to their father and mother. Every agent has an attribute called _fatherId_, and I would need to do something like: `import networkx as nx G=nx.Graph() for person in entries: G.add_edge(person, person_with_id_equal_to_fatherId)` – Wilco May 16 '12 at 17:50
Yep, probably you should create another post where you explain the last part of your problem more deeply. You probably have a way to do the matching by id by implementing a dictionary functionality in your class but as I said it would be great if you could explain that part of the problem in a different post. – txominpelu May 17 '12 at 11:36

score 1 · Answer 2 · answered May 14 '12 at 12:56

I have a .csv file

You're in luck; CSV support is built right in, via the csv module.

Do you suggest creating a class dictionary for accessing every instance?

I don't know what you think you mean by "class dictionary". There are classes, and there are dictionaries.

But I still need to provide a name to every single instance, right? How can I do that dynamically? Probably the best thing would be to use the unique ID, read from the file, as the instance name, but I think that numbers can't be used as instance names, can they?

Numbers can't be instance names, but they certainly can be dictionary keys.

You don't want to create "instance names" dynamically anyway (assuming you're thinking of having each in a separate variable or something gross like that). You want a dictionary. So just let the IDs be keys.

I miss pointers! :(

I really, honestly, can't imagine how you expect pointers to help here, and I have many years of experience with C++.

Thanks for your answer. By class dictionary I mean a dictionary which is an attruibute of a class. I don't know if this makes sense. Actually I DO want to create instances automatically from every csv line, because I want each person to be an instance of the class Person(). I'll use the tools suggested by user433831, but still I don't want to do that manually! As for pointers, I mean that in C++ you don't have to give a name to a class instance, you just use the "new" operator. — Wilco, May 15 '12 at 16:33
You don't have to give names to in Python, either. If you have `class Foo: ...`, then `Foo()` is a perfectly valid line of Python code by itself. Or you could have, for example, `mylist = []; mylist.append(Foo())`. — Karl Knechtel, May 16 '12 at 03:16

score 0 · Answer 3 · answered May 14 '12 at 12:36

good day. you may make it like this:

    >>> class person(object):
    ...     def __init__(self, name):
    ...             self.name = name
    ...
    >>> d = dict()
    >>> names = ['1','2','3','4','5']
    #fill dict with any value. for example with hash of name        
        >>> for name in names:      
    ...     d[person(name)] = hash(name)
    ...
    >>>
    >>> d
    {<__main__.person object at 0x000000000236BC18>: 1977051568, <__main__.person ob
    ject at 0x0000000002372198>: -1933914571, <__main__.person object at 0x000000000
    1E95FD0>: 2105051955, <__main__.person object at 0x0000000002372160>: -180591418
    8, <__main__.person object at 0x00000000023721D0>: -2061914958}
    >>>

    >>> for key in d.keys():
    ...     key.name
    ...
    '1'
    '4'
    '2'
    '5'
    '3'

Now We have 5 objects.

How can I dynamically generate class instances with single attributes read from flat file in Python?

3 Answers3

Linked