1 00:00:05,280 --> 00:00:10,309 We've learned a lot about Python dictionaries. Now let's see how the documentation describes them. 2 00:00:10,309 --> 00:00:14,640 I'm going to go to a link, to the python documentation on screen. 3 00:00:14,640 --> 00:00:20,240 As always, the link is in the resources section. This is a more formal description of the dict type - 4 00:00:20,240 --> 00:00:26,781 and it looks scary! But don't worry - the more you work with this documentation, the more sense it'll make. 5 00:00:26,781 --> 00:00:30,960 So starting from the top, the documentation talks about Mapping Types, and notes that there's 6 00:00:30,960 --> 00:00:36,880 currently only one mapping type built into Python. That's the dict type that we've been using. 7 00:00:36,880 --> 00:00:41,760 Those hashable values, in the first paragraph, are the keys. The documentation tries to be very 8 00:00:41,760 --> 00:00:48,000 concise, and accurate. It talks about hashable values, rather than keys, to document exactly 9 00:00:48,000 --> 00:00:54,800 what can be used as keys - and that'll make more sense soon. It also mentions arbitrary objects. 10 00:00:54,800 --> 00:01:00,480 It uses that phrase because you can store anything in a dictionary. The keys must be hashable values, 11 00:01:00,480 --> 00:01:06,320 whatever they are, but the values we store in our dictionaries can be any Python object. You can even 12 00:01:06,320 --> 00:01:12,330 store a dictionary or a list inside another dictionary. The second paragraph says that 13 00:01:12,330 --> 00:01:19,360 A dictionary's keys are almost arbitrary values. What that means is, you can only use immutable objects 14 00:01:19,360 --> 00:01:25,600 as a key. It goes on to clarify that, saying that mutable types may not be used as keys. 15 00:01:25,600 --> 00:01:31,280 ints and strings are immutable, and are commonly used as dictionary keys. But if you need to, 16 00:01:31,280 --> 00:01:37,520 you could use a tuple as a key. Let's click on one of these hashable links and open it up in a new tab, 17 00:01:37,520 --> 00:01:41,840 to see what the documentation means by hashable. 18 00:01:41,840 --> 00:01:48,000 Once again, the text here is very concise and precise. A hash value is a number that's calculated 19 00:01:48,000 --> 00:01:53,840 from an object's contents. For integers, the hash value is the integer itself. For things like 20 00:01:53,840 --> 00:01:59,600 strings, a special function is used to calculate the numerical hash value. We'll see why hashes are 21 00:01:59,600 --> 00:02:06,160 useful, and how a dictionary uses them, when we get to the theory behind mappings. The third paragraph 22 00:02:06,160 --> 00:02:13,220 clarifies what it means for an object to be hashable. Hashable objects must be immutable. 23 00:02:13,680 --> 00:02:18,720 But it does match some restrictions. We know that tuples are immutable, so they could be 24 00:02:18,720 --> 00:02:25,186 used as a dictionary key. But if a tuple contains a list, then it's no longer suitable as a key. 25 00:02:25,186 --> 00:02:33,230 So a tuple can be used as a dictionary key, as long as the items in the tuple are also immutable. 26 00:02:33,680 --> 00:02:37,840 Looking at this table, the left-hand column shows some immutable objects that we can use 27 00:02:37,840 --> 00:02:44,400 as dictionary keys. The first item is a tuple that only contains other immutable items. 28 00:02:44,400 --> 00:02:50,480 The tuple on the right hand side isn't suitable. It contains a list, and the list could be mutated. 29 00:02:50,480 --> 00:02:56,000 Our code could add items to it, or delete them. Because of that, we can't use this tuple as a 30 00:02:56,000 --> 00:03:01,600 dictionary key. Python won't attempt to create a hash for a tuple that contains things like lists. 31 00:03:01,600 --> 00:03:08,006 You'll get a TypeError: unhashhable type if you attempt to hash the tuple on the right. 32 00:03:08,720 --> 00:03:14,240 Alright, so let's go back to the documentation about the dict mapping type. 33 00:03:14,240 --> 00:03:20,400 So the next bit of documentation - scroll down a bit further - shows various ways to use the 34 00:03:20,400 --> 00:03:25,600 dict constructor. Now don't worry about what a constructor is. For now, just treat it like 35 00:03:25,600 --> 00:03:30,240 another function. You've seen dict in the list of Python built-in functions that we've looked at, 36 00:03:30,240 --> 00:03:36,800 a few times. If we scroll down a bit further, the examples down here in this code block, all show 37 00:03:36,800 --> 00:03:43,760 different ways to create a dictionary with the same keys and values; b uses a dict literal, and is 38 00:03:43,760 --> 00:03:48,880 the way that we've been creating our dictionaries so far. The documentation example shows the code 39 00:03:48,880 --> 00:03:55,280 on a single line, but in our examples, we split our literals over several lines. d and e show how to 40 00:03:55,280 --> 00:04:00,560 pass a list of tuples, or another dictionary, to the dict function. You probably won't write code 41 00:04:00,560 --> 00:04:06,560 like the example for e, but you could pass b to the dict function. That would create a new dictionary, 42 00:04:06,560 --> 00:04:12,320 with the same values as b. I'll skip the examples that use keyword arguments - that's the lines that 43 00:04:12,320 --> 00:04:17,519 create a and f. We haven't covered keyword arguments yet. If you remember, I said you'd 44 00:04:17,519 --> 00:04:22,160 need to know about dictionaries before you could understand how keyword arguments work. But if 45 00:04:22,160 --> 00:04:27,120 you've guessed that keyword arguments are stored in a dictionary, you're right. Below that, as we go 46 00:04:27,120 --> 00:04:32,320 through the remainder of the section, it describes the operations that a dictionary supports. We've 47 00:04:32,320 --> 00:04:36,320 actually used a lot of these already, and I'm going to discuss some of the others that aren't obvious. 48 00:04:36,320 --> 00:04:41,200 I don't think I need to explain what len is, or scrolling down a little bit further, 49 00:04:41,200 --> 00:04:48,800 what clear is. Going back up here a little bit, list and len are both Python built-in functions. 50 00:04:48,800 --> 00:04:53,520 That's why they appear before the alphabetical list of operations and methods. If we scroll a 51 00:04:53,520 --> 00:05:00,080 little bit further, down here, there's some examples of accessing items using a key, setting a value - 52 00:05:00,080 --> 00:05:05,680 again, using a key - and removing an item from the dictionary. The documentation uses the letter d to 53 00:05:05,680 --> 00:05:10,800 represent a dictionary, in those examples. The first function we haven't used so far, is iter. 54 00:05:10,800 --> 00:05:15,520 Using iterators, explicitly, is quite advanced. You've been using an iterator every time you 55 00:05:15,520 --> 00:05:21,600 use a for loop. An iterator is created implicitly, by the for statement. Later in the course, we'll see 56 00:05:21,600 --> 00:05:26,080 some examples of why you might want to create your own iterator. But it's not something you need to do 57 00:05:26,080 --> 00:05:31,760 very often, and I won't discuss it just yet. Like list and len, iter is a built-in function that can 58 00:05:31,760 --> 00:05:37,120 be used with any iterable or sequence type. It's not specific to dictionaries - you can pass a list 59 00:05:37,120 --> 00:05:43,120 or a string to those three functions. We're getting now, to the methods that we can use with a dict. 60 00:05:43,120 --> 00:05:47,564 Dot clear shouldn't need any explanation - it clears all the items out of a dictionary. 61 00:05:47,564 --> 00:05:52,480 The copy method creates a copy of a dictionary. That's fairly straightforward, but why does it 62 00:05:52,480 --> 00:05:57,280 mention a shallow copy? Do we get a dictionary that's only interested in your looks? Or in how 63 00:05:57,280 --> 00:06:02,480 much you earn? Of course not. The difference between a shallow copy and a deep copy is quite 64 00:06:02,480 --> 00:06:07,600 important, and we'll discuss that in the next few videos. I'll also give some examples of 65 00:06:07,600 --> 00:06:14,640 some of the other methods that we haven't used yet. I won't provide an example using popitem, 66 00:06:14,640 --> 00:06:19,440 because it's very similar to pop. Whereas pop returns the value from the dictionary, popitem 67 00:06:19,440 --> 00:06:24,320 returns both the key and the value, as a tuple. There are a couple of things to be aware of with 68 00:06:24,320 --> 00:06:30,480 popitem. The first is, you don't tell it which key to remove - it automatically removes the last item 69 00:06:30,480 --> 00:06:39,760 that was added to the dictionary. That's what LIFO means - it's an acronym meaning Last In, First Out. 70 00:06:39,760 --> 00:06:44,960 Imagine someone wearing bangles on their wrist. If you put on a red bangle, a blue bangle and a green 71 00:06:44,960 --> 00:06:50,939 bangle, then they'll take them off in the opposite order. The green bangle has to come off first - last on, 72 00:06:50,939 --> 00:06:56,000 first off. Another example would be a stack of coins. The first coin you take off the stack would 73 00:06:56,000 --> 00:07:02,903 be the last one you put on it. In fact, a stack is a very common data structure in programming. 74 00:07:04,160 --> 00:07:09,360 The opposite of LIFO is FIFO - First In, First Out. that's like a queue at 75 00:07:09,360 --> 00:07:14,480 a bus stop. The first person to arrive at a bus stop will be the first person to leave, 76 00:07:14,480 --> 00:07:19,926 when they get on the bus. We're assuming a civilized behaviour here, of course. 77 00:07:20,800 --> 00:07:26,000 Another example of a FIFO queue would be putting items away in your fridge. If you bought more 78 00:07:26,000 --> 00:07:31,360 milk, then you'd put the new bottle behind any that were already in the fridge. When you take 79 00:07:31,360 --> 00:07:35,600 a bottle of milk out of the fridge, then you'd be taking the first bottle that was put in. 80 00:07:35,600 --> 00:07:41,661 If you don't do that, you could end up with a bottle of milk that's been in the fridge for months. Yeuch! 81 00:07:42,000 --> 00:07:47,520 The other thing to watch out for with popitem, is that you only get LIFO order with Python 82 00:07:47,520 --> 00:07:53,840 version 3.7 and higher. Python versions before 3.7 will return and remove random items from 83 00:07:53,840 --> 00:07:58,400 the dictionary. So pay careful attention to notes like this in the documentation. 84 00:07:58,400 --> 00:08:03,440 If you write a program that relies on popping the last inserted items, your code won't work 85 00:08:03,440 --> 00:08:08,720 on earlier versions of Python. Looking down a little bit further, we've used the reversed function 86 00:08:08,720 --> 00:08:12,640 when working with lists, and it does the same thing with a dictionary. It lets you iterate 87 00:08:12,640 --> 00:08:18,400 backwards over the dictionary. Now note that it won't work with Python versions before 3.8. 88 00:08:18,400 --> 00:08:21,600 A requirement to iterate backwards is that the keys are ordered, 89 00:08:21,600 --> 00:08:28,160 and that change was made in python 3.7. Python 3.8 could take advantage of that, to allow reversed to 90 00:08:28,160 --> 00:08:32,159 work with a dictionary. Alright, so I'll stop the video here, and in the next one, we'll look 91 00:08:32,159 --> 00:08:37,840 at methods I've skipped over, and see some short examples of each one. See you in the next video.