I've noticed when creating a large amount (around 25000) of new ORM objects in a loop, (in this case Treebeard MP_nodes) Django begins eating up more and more of the machines ram in gigabytes as it iterates through an in-memory dictionary and creates these models, and i can't wrap my head around why it does so.

​

I've run into this before when creating objects in a migration, so i don't believe its specific to the Treebeard library. I was wondering if any of you guys have run into the same issue when creating a large amount of models without bulk create.

​

Obviously when doing this you should use bulk_create, But this is a limitation of the library which doesn't have a method to bulk create the nodes of the tree at once. I can get around this by using some private methods, but I'd rather not unless I have to.