Question

I am looking for the best way to create a list in python that creates hashed indexes (dicts) for all the properties of the objects put into the list.

>>> foo = IndexingList([{ 'id': 1, 'name': 'cat' }, { 'id': 2, 'name': 'dog' }])
>>> foo[0]
{'id': 1, 'name': 'cat'}

>>> foo.findall('id', 2)
[{'id': 2, 'name': 'dog'}]

>>> foo += {'id': 3, 'name': 'dog'}
>>> foo.findall('name', 'dog')
[{'id': 2, 'name': 'dog'}, {'id': 3, 'name': 'dog'}]

I imagine the data structure of the IndexingList would then look like this:

{
    'items': [
        { 'id': 1, 'name': 'cat' }, 
        { 'id': 2, 'name': 'dog' }
    ],
    'indexes': {
        'id': {
            1: [{ 'id': 1, 'name': 'cat' }],
            2: [{ 'id': 2, 'name': 'dog' }]
        },
        'name': {
            'cat': [{ 'id': 1, 'name': 'cat' }],
            'dog': [
                { 'id': 2, 'name': 'dog' },
                { 'id': 3, 'name': 'dog' }
            ]
        }
    }
}

where the objects within the ‘indexes’ nodes refer to the same objects in ‘items’.

I think property values that are themselves objects could receive unique index-keys by using str(property) to obtain something to stick in ‘indexes’.

Answer 1

This is actually pretty easy to do using some collections.defaultdict()s – although you might consider using an actual database if you are using this a lot.

from collections import defaultdict
from functools import partial

class IndexingList:
    def __init__(self, items):
        self.items = []
        self.indices = defaultdict(partial(defaultdict, list))
        self.extend(items)

    def append(self, item):
        try:
            for index, value in item.items():
                self.indices[index][value].append(item)
        except AttributeError as e:
            raise ValueError("All children of an IndexingList must be "
                             "dict-like. '{0}' is not.".format(item)) from e
        self.items.append(item)

    def extend(self, iterable):
        for item in iterable:
            self.append(item)

    def __iadd__(self, other):
        self.extend(other)
        return self

    def __getitem__(self, item):
        return self.items[item]

    def __setitem__(self, item, value):
        self.items[item] = value

    def __delitem__(self, item):
        del self.items[item]
        for index, value in item.items():
            self.indices[index][value].remove(item)

    def find_all(self, index, value):
        return self.indices[index][value]

    def __repr__(self):
        return repr(self.items)

Used like so:

>>> foo = IndexingList([{ 'id': 1, 'name': 'cat' }, { 'id': 2, 'name': 'dog' }])
>>> foo[0]
{'id': 1, 'name': 'cat'}
>>> foo.find_all("id", 2)
[{'id': 2, 'name': 'dog'}]
>>> foo += [{'id': 3, 'name': 'dog'}]
>>> foo.find_all('name', 'dog')
[{'id': 2, 'name': 'dog'}, {'id': 3, 'name': 'dog'}]