1 00:00:05,280 --> 00:00:10,240 I'll finish our discussion of deep and shallow copies with a quick example. We'll perform a deep 2 00:00:10,240 --> 00:00:15,040 copy of our dictionary and examine the difference. What we need to do is open up our earlier 3 00:00:15,040 --> 00:00:21,920 shallow_copy program, and we're going to change this to be using a deep copy instead. 4 00:00:21,920 --> 00:00:27,678 First though, a quick review. We'll run the program, then we'll check the output. 5 00:00:27,678 --> 00:00:34,378 On line 14, we use the reference to the teddy list in the things dictionary. We append toy to the list. 6 00:00:34,378 --> 00:00:39,237 The last two lines of the output are the same - the toy entry appears in both lines of output. 7 00:00:39,237 --> 00:00:42,960 Now this should no longer be surprising. The value in both dictionaries 8 00:00:42,960 --> 00:00:48,480 are referring to the same list. Now let's have a look at a deep copy. As you've seen 9 00:00:48,480 --> 00:00:54,720 from the documentation that I linked to, in the last video, the copy module provides a deep copy function. 10 00:00:54,720 --> 00:01:01,156 To use the function, we need to import it into our program. Going up to line one, 11 00:01:02,640 --> 00:01:07,840 import copy. We'll be looking at importing in more detail, because there are different 12 00:01:07,840 --> 00:01:13,027 ways to import things, but at the moment we're focusing on dictionaries and copying. 13 00:01:13,027 --> 00:01:17,520 The import's very similar to when we imported the random module, back in the Program Flow Control and 14 00:01:17,520 --> 00:01:23,040 Python section. That was in the video The Random Module and Import. So review that video, 15 00:01:23,040 --> 00:01:27,120 if you want to remind yourself what's happening here. Alright, so let's see what happens when we perform 16 00:01:27,120 --> 00:01:32,080 a deep copy, rather than a shallow copy. So I'm going to add some comments to the code, comment 17 00:01:32,080 --> 00:01:38,160 out something and then use the deep copy function. I'm going to come down here and comment out line 9. 18 00:01:38,160 --> 00:01:45,920 Line 8, which will become line 9, I'm going to comment Perform a shallow copy, 19 00:01:45,920 --> 00:01:53,802 because that's obviously what that was doing. And down here, comment Perform a deep copy. 20 00:01:53,802 --> 00:02:02,472 Then the actual code will be things equals copy dot deepcopy, and in parentheses, animals. 21 00:02:03,200 --> 00:02:08,767 So this time when I run this, we can see that toy is only added to the list in the things dictionary. 22 00:02:08,767 --> 00:02:15,064 Each dictionary now contains its own copy of the lists. When we append to the list in things, on line 20, 23 00:02:15,064 --> 00:02:18,880 it's got no effect on the list in animals - we can see that with the output. So it might actually 24 00:02:18,880 --> 00:02:25,864 help to confirm that, by printing out the ids of the list. Alright, so I'm going to copy that part of the printout. 25 00:02:25,864 --> 00:02:30,434 I'm going to add something at the start: id parentheses, paste that in, and put a 26 00:02:30,480 --> 00:02:34,960 comma outside of the right parentheses and do the same for the next line as well. So we're printing 27 00:02:34,960 --> 00:02:46,735 out the id as well as the actual values. So id parentheses, paste it in, and a comma. We'll run this 28 00:02:47,120 --> 00:02:51,520 So the first two lines of output, produced by lines 15 and 16, 29 00:02:51,520 --> 00:02:55,760 show that the lists have different ids. They're not the same list, and it's not surprising 30 00:02:55,760 --> 00:03:01,760 that toy isn't added to both lists. Now if we swap the commenting around, 31 00:03:01,760 --> 00:03:09,680 by uncommenting line 10 and commenting out line 13, we'll go back to performing a shallow copy. 32 00:03:09,680 --> 00:03:17,600 We run this. This time you can see the ids are the same. Both dictionaries contain references to the 33 00:03:17,600 --> 00:03:24,240 same lists. Because there's only one list, we get toy appearing in both the last two lines of output. 34 00:03:24,240 --> 00:03:28,287 Okay, so that's the difference between deep and shallow copies. 35 00:03:28,287 --> 00:03:33,259 A deep copy will also copy any objects that are contained in whatever you're copying. 36 00:03:33,259 --> 00:03:39,360 A shallow copy only copies references - it doesn't make copies of contained objects. I'm going to 37 00:03:39,360 --> 00:03:43,040 finish our discussion of dictionaries with some theory. We'll look at how dictionaries 38 00:03:43,040 --> 00:03:50,917 manage to be so fast when retrieving items using a key. But before then, it's time for a challenge. 39 00:03:51,600 --> 00:03:55,613 Your challenge is to write a function that takes the dictionary as an argument, 40 00:03:55,613 --> 00:04:00,640 and returns a deep copy of the dictionary. You're going to write your own function, to do a similar 41 00:04:00,640 --> 00:04:06,720 job to the deepcopy function that we've just used, but you'll do it without using the copy module. 42 00:04:06,720 --> 00:04:10,800 Your function will be a lot simpler than deepcopy. It only has to cope with 43 00:04:10,800 --> 00:04:16,399 one level of contained objects. It should be able to successfully copy dictionaries like our animals 44 00:04:16,399 --> 00:04:20,320 or recipes dictionaries. It doesn't have to handle dictionaries 45 00:04:20,320 --> 00:04:25,892 that contain objects that also contain objects. That's much too difficult at this stage. 46 00:04:26,960 --> 00:04:31,040 The basic approach will be to create a new empty dictionary. 47 00:04:31,040 --> 00:04:35,380 Next, iterate over the keys and values of the dictionary that's being copied. 48 00:04:35,380 --> 00:04:41,040 For each key, copy the value, then add the copy of the value to the new dictionary. Your function 49 00:04:41,040 --> 00:04:47,021 only has to handle values that are dictionaries or lists. Both of those objects have a copy method. 50 00:04:48,400 --> 00:04:53,008 The code to test your function might look something like you can see on screen here, 51 00:04:53,008 --> 00:04:55,840 so pause the video if you want to look at that. 52 00:05:00,160 --> 00:05:03,422 And this is the output you should get. 53 00:05:04,400 --> 00:05:06,720 Alright, so that's the challenge, see how you go with it, 54 00:05:06,720 --> 00:05:09,840 and I'll go over my solution to it in the next video.