Is there a more efficient way to allocate lots of instances of a class in a loop?

Question

I am working on a Python program, and there is an inner loop which is causing performance problems. I ran a profiler on it, and I found that 50% of the time is spent in object allocation. Basically, something like this:

result_list = []
for i in range(BIG_NUMBER):
    obj = MyClass()
    # do stuff with obj...
    result_list.append(obj)

The line obj = MyClass() takes 50% of the time. Is there any way to efficiently allocate lots of these objects in Python? Perhaps some sort of bulk allocation?

Without copy / pasting the entire MyClass, it has the following characteristics:

• It is an ORM object, so it basically maps attributes to database values
• It has quite a few property functions, a meta class, etc. so it's a pretty "fat" class

The # do stuff with obj... line largely sets the various attributes, usually via setattr, but in a few cases via a custom function. So if there are 10 attributes to set, 9 will usually be via setattr, and 1 might be via a special function. This part takes ~40% of the time and I'm optimizing it separately.

Answer 1

If the object starts w an empty constructor call, could you not pull it out if the loop? Then copy.deepcopy your seed object?