1 00:00:05,280 --> 00:00:09,360 Alright, so you should now be comfortable returning values from a dictionary, using 2 00:00:09,360 --> 00:00:14,800 their keys. Let's see a few more things we can do with dictionaries. Although they're not a 3 00:00:14,800 --> 00:00:21,040 sequence type, they are iterable. We can iterate over the keys in a dictionary. Now the syntax 4 00:00:21,040 --> 00:00:25,440 is the same as we've used before, so let's have a look at the code. So what I'm going to do is 5 00:00:25,440 --> 00:00:31,920 delete the code from the last video, and just keep the dictionary definition. Alright, so I'm going 6 00:00:31,920 --> 00:00:39,840 to start by typing for key in vehicles colon, and print key in parentheses on the next line. 7 00:00:42,640 --> 00:00:43,840 Alright, so if we run this now, 8 00:00:49,200 --> 00:00:52,320 you can see there, that we've got the dictionary keys printed out. 9 00:00:52,320 --> 00:00:58,400 So when you iterate over a dictionary, you're iterating over the keys. If you want the values, 10 00:00:58,400 --> 00:01:04,160 you might decide to index the dictionary. I'll do that to print out the values, as well as the keys. 11 00:01:04,160 --> 00:01:10,560 So I'm going to change line 13, specifically, and put comma after the key and vehicles, 12 00:01:11,440 --> 00:01:17,200 then left square bracket, key. So key in square brackets, in other words. 13 00:01:18,547 --> 00:01:25,175 Alright, so if we run this now, you can see we've got the key this time, followed by the value. 14 00:01:25,175 --> 00:01:31,704 But it's not obvious which is which here, but we can fix that by using f strings to format the output better. 15 00:01:31,704 --> 00:01:34,880 We'll get plenty of practice doing that, so I'll do something different here. 16 00:01:35,600 --> 00:01:40,720 If you recall, we saw the sep argument that we can pass to the print function, back in section five, 17 00:01:40,720 --> 00:01:46,640 but you may have forgotten about it by now. So what I'm going to do is use it here, as a reminder - 18 00:01:46,640 --> 00:01:52,085 just as a reminder that it's there if you need it. What I'm going to do after the right square bracket there, 19 00:01:52,085 --> 00:02:00,000 is put comma sep equals, and in double quotes, a comma and a space. Run the program again now. 20 00:02:01,120 --> 00:02:06,480 This time you can see that we've got the comma and a space between the key and the value itself. 21 00:02:07,600 --> 00:02:13,360 So it's more efficient to use the items of a dictionary instead. Back in the Lists, Ranges 22 00:02:13,360 --> 00:02:19,520 and Tuples section, we talked about the enumerate function. That provides an efficient way to get the 23 00:02:19,520 --> 00:02:25,760 indexes and items from a sequence. If you remember, we unpacked the index and value from the tuples 24 00:02:25,760 --> 00:02:31,200 that enumerate returns. If you don't remember, review the lecture called The enumerate Function 25 00:02:31,200 --> 00:02:36,800 in that section, and the Improving our Code lecture that followed it. Now we can do something similar 26 00:02:36,800 --> 00:02:42,480 with dictionaries. We don't have to call enumerate, because dictionaries have the dot items method 27 00:02:42,480 --> 00:02:47,920 that we can use instead. So let's change that code to do the same thing more efficiently. 28 00:02:47,920 --> 00:02:52,171 So what I'm going to do is comment out the current loop and just add a new one. 29 00:02:55,600 --> 00:02:59,040 Alright, so I'm going to type for key comma value 30 00:03:00,000 --> 00:03:08,085 in vehicles dot items, and you can see there with parentheses on the end, colon. Then we're going to print, 31 00:03:08,085 --> 00:03:15,330 and in parentheses, key comma value comma sep equals and in double quotes, comma space. 32 00:03:17,200 --> 00:03:23,769 So if we run the program now, you can see we've got the same output as we got previously. 33 00:03:25,120 --> 00:03:28,480 So students sometimes object when we show code that you shouldn't write. 34 00:03:29,120 --> 00:03:34,560 We've done it because the loop on lines 12 and 13 - the loop that I've now commented out - is the 35 00:03:34,560 --> 00:03:39,680 way most new programmers would try to do this. It works, so they develop the habit of writing their 36 00:03:39,680 --> 00:03:45,120 code like that. Now it's not bad, but it's less efficient than iterating over items like we're 37 00:03:45,120 --> 00:03:52,658 doing, on lines 14 and 15. With larger data sets, those few milliseconds can soon add up. 38 00:03:53,840 --> 00:03:57,280 you might come across that inefficient loop if you find yourself modifying 39 00:03:57,280 --> 00:04:04,160 or converting old python 2 code. dot items in python 2 created a list. That means it needed a 40 00:04:04,160 --> 00:04:09,920 copy of the data, and used more memory. For that reason, programs might have used the simple for 41 00:04:09,920 --> 00:04:18,320 loop to avoid the extra memory overhead of using dot items. Python 2.2 added a dot iteritems method 42 00:04:18,320 --> 00:04:24,560 which works in a similar way to python 3's dot items method. Python 3 now uses something called 43 00:04:24,560 --> 00:04:30,960 a generator, and doesn't copy the data from the dictionary. With python 3, remember to use enumerate 44 00:04:30,960 --> 00:04:36,739 when iterating over sequences, and dot items when iterating over dictionaries. 45 00:04:37,276 --> 00:04:44,560 Alright, so the dot items method returns an iterator of tuples, and we unpack the tuples into our key and value variables, 46 00:04:44,560 --> 00:04:51,120 on line 14. Alright, so that's how we can iterate over the keys and items of a dictionary. 47 00:04:51,120 --> 00:04:56,351 I'll finish the video with a bit of history, because some students might be getting strange results. 48 00:04:57,520 --> 00:05:02,560 If you installed Python from the python.org website, the implementation of Python that you 49 00:05:02,560 --> 00:05:08,400 installed from that site is called CPython. That's because most of the implementation is written in 50 00:05:08,400 --> 00:05:14,480 C. There are other implementations of Python, for example Iron Python, or IPython, and Jython. 51 00:05:16,960 --> 00:05:22,080 CPython is the reference implementation of Python, the version written by Guido 52 00:05:22,080 --> 00:05:27,920 van Rossum. Version 3.6 of CPython changed the way that dictionaries worked. 53 00:05:27,920 --> 00:05:32,320 In earlier versions, the keys of a dictionary were unordered. That means you could get the 54 00:05:32,320 --> 00:05:38,320 keys printed in a different order each time you ran the program. If you're using python 3.5, 55 00:05:38,320 --> 00:05:42,000 and the dictionary prints in a different order each time, now you know why that is. 56 00:05:43,920 --> 00:05:49,200 Python 3.6 preserved the insertion order of dictionary keys. That means you'll always 57 00:05:49,200 --> 00:05:54,720 iterate over them in the same order. That was an implementation detail, which meant you couldn't 58 00:05:54,720 --> 00:06:02,000 rely on the behavior in other implementations of python 3.6. With python 3.7, that behavior 59 00:06:02,000 --> 00:06:07,200 became part of the language. When you iterate over a dictionary, with python 3.7 and above, 60 00:06:07,200 --> 00:06:13,120 the keys will appear in the order they were added to the dictionary. Now that's something you need 61 00:06:13,120 --> 00:06:19,040 to be aware of, because you can get a different order in earlier versions of Python. Python 3.5 62 00:06:19,040 --> 00:06:24,160 goes out of support in September 2020 but that doesn't mean people will stop using it. But it 63 00:06:24,160 --> 00:06:33,112 does mean it's acceptable for you to specify that your code requires pythons 3.6 or higher, if you rely on features introduced in that version. 64 00:06:33,335 --> 00:06:37,920 Alrighty, so let's move on now to the next video.