python设置与对象集的交集

Question

I am working with amazon boto and I have 2 lists. List 1 contains Instance objects. List 2 contains InstanceInfo objects. Both objects have a attribute called id. I need to get a list of Instance objects which id exists in the InstanceInfo list.

l1 = [Instance:i-04072534, Instance:i-06072536, Instance:i-08072538, Instance:i-0a07253a, Instance:i-e68fa1d6, Instance:i-e88fa1d8, Instance:i-ea8fa1da, Instance:i-ec8fa1dc]

l2 = [InstanceInfo:i-ec8fa1dc, InstanceInfo:i-ea8fa1da, InstanceInfo:i-e88fa1d8, InstanceInfo:i-e68fa1d6]

Wanted result:

l3 = [Instance:i-ec8fa1dc, Instance:i-ea8fa1da, Instance:i-e88fa1d8, Instance:i-e68fa1d6]

Right now I have it working through:

l3= []
for a in l1  
    for b in l2:
        if a.id == b.id:
            l3.append(a)

However, I have been told that I should replace this using set intersection. I have been looking at examples and it looks very straightforward. Yet I don't see any examples working with objects.

I've been playing around for a bit and theoretically I can see it work, yet there might be some 'advanced' syntax that I may not know off. I am still learning python.

Answer 1

Here is something faster than Marcin's answer (while being similar):

ids_l1 = set(x.id for x in l1)  # All ids in list 1
intersection = [item for item in l2 if item.id in ids_l1]  # Only those elements of l2 with an id in l1

It is important to pre-calculate ids_l1 and to not write if item.id in set(…) , as the set would be reconstructed each time (as the full test expression is re-evaluated for each element item ).

Python sets give you fast element membership tests ( in ). Such tests are much faster with sets than with lists (as the elements of a list must be read one by one, whereas the elements of a set are "hashed").

Answer 2

Your method may be relatively efficient for smallish lists.

With sets, you would have to extract the ids, calculate the intersection of ids, and then collect the items into your new list. Something like:

set1 = set(x.id for x in l1)
set2 = set(x.id for x in l2)
intersection_ids = set1 & set2
intersection_list = [item for item in l2 if item.id in intersection_ids]

You can make this a little more efficient by scanning over the shorter list, or by storing your objects in a dict.

Answer 3

Try this:

# get ids of elements in second list
l2_ids = [x.id for x in l2]
# get elements from first list that have ids in second
l3 = [x for x in l1 if x.id in l2_ids]