WEBVTT 00:02.960 --> 00:07.000 Hello again! In this video, we're going to look at numeric types and literals. 00:09.760 --> 00:14.080 In C++, the size of the built in types isn't precisely defined. 00:14.470 --> 00:16.000 It depends on the implementation. 00:17.740 --> 00:22.750 The only one that is fixed is char, which has to be the byte size of the machine. And on anything that 00:22.750 --> 00:24.820 you likely to use, that's going to be 8 bits. 00:25.870 --> 00:30.550 An int has to be at least 16 bits. A long has to be at least 32 bits, 00:30.760 --> 00:37.150 and at least as many as int. And C++ 11 has a long long type, which has to be at least 64 bits, 00:37.750 --> 00:39.760 and it must be at least as many as long. 00:41.900 --> 00:48.830 If you want a fixed-width integer which has a well-defined size, then we have these now in the library. 00:49.010 --> 00:51.110 So you need to include the C standard int header. 00:51.680 --> 00:53.270 These have actually been brought in from C. 00:53.630 --> 00:57.620 So we have types for eight, 16, 32 and 64 bit ints. 00:57.950 --> 00:59.780 And also the unsigned versions. 01:04.720 --> 01:05.890 So let's see how these work. 01:10.940 --> 01:17.840 So there we are. char is one byte, which it is by definition. int is four bytes on this machine and long 01:17.840 --> 01:19.910 is also four bytes. Long 01:19.910 --> 01:23.810 long is eight bytes, which is 64 bits, so that's the minimum it could be. 01:24.530 --> 01:28.640 And then the fixed width types are what we would expect them to be. 01:35.670 --> 01:40.530 We can also use different bases in C++. If we don't say which base we want, then it's assumed we 01:40.530 --> 01:41.280 want decimal. 01:41.280 --> 01:44.700 So 42 is the number 42 in base. 01:44.700 --> 01:45.000 10. 01:46.110 --> 01:51.600 If we put 0x in front - it can be upper or lowercase - then that's interpreted as hexadecimal. 01:52.050 --> 01:56.460 So zero x 2 a that's 42 in base 16. 01:58.230 --> 02:01.380 There's also octal, which is base eight. 02:01.380 --> 02:05.490 So that has a leading zero. And there is actually a bit of a trap here. 02:05.790 --> 02:11.580 If you are dealing with strings that contain numbers, possibly with a GUI, then they may have leading 02:11.580 --> 02:12.120 zeroes. 02:12.480 --> 02:16.440 And in that case, when you convert them to integers, they are going to be interpreted as octal. 02:17.040 --> 02:21.760 So if you have something that starts with zero eight or zero nine, that's an error because octal only 02:21.760 --> 02:22.500 goes up to seven. 02:22.920 --> 02:27.810 And if it starts with zero one, zero two, zero seven, then you just get the wrong number. 02:29.520 --> 02:30.960 So that's something to watch out for. 02:31.080 --> 02:32.550 I have actually been caught by that. 02:34.870 --> 02:41.920 And finally, in C++ 14, we can also have binary literals with zero b and that's 42 in 02:41.920 --> 02:42.100 binary. 02:43.390 --> 02:44.890 So let's just try that out. 02:45.310 --> 02:49.780 So there we are. OK, they all give 42 in different bases. 02:55.040 --> 02:59.570 Floating points. Again, these aren't precisely defined, but you can be fairly certain what you get 02:59.570 --> 03:05.060 on current hardware. With floats, you will usually get six digits precision, so that's all the digits 03:05.060 --> 03:05.600 in the number. 03:06.140 --> 03:12.710 Regardless of whether they are before or after the decimal point. With double, we get 15 digits and long 03:12.710 --> 03:16.640 double is 20, which came in in C++ 98, I think. 03:19.850 --> 03:25.310 If you have a number with a lot of digits in, you can put a separator in to make it easier to read. That's 03:25.310 --> 03:29.950 just the quote character, so we can write one million with the quote inside it. 03:30.380 --> 03:32.450 And this came in in C++ 14. 03:33.560 --> 03:35.120 And it works with decimals as well. 03:38.510 --> 03:43.550 Normally, you'll probably use the Western convention, in which there's a separator after every three digits, 03:43.550 --> 03:47.390 but there are other conventions. And you can in fact put the separator anywhere you want. 03:47.780 --> 03:53.630 So if for example you want to use the Indian convention, then you can just put the separator in the appropriate 03:53.630 --> 03:54.020 place. 03:55.040 --> 03:59.810 The separator has to go inside the number. It can't be at the beginning or the end as otherwise that will 03:59.810 --> 04:01.760 be interpreted as a character literal. 04:05.310 --> 04:07.890 OK, so here's some code, so we can just try this out. 04:07.920 --> 04:13.980 We have one million, one lakh and we have pi, using these separators, and then we're going to print them out. 04:14.870 --> 04:17.790 And yep, so that's one million. 04:17.970 --> 04:23.130 It's quite hard to tell just from looking at that, that is just one million and not 100,000. 04:23.940 --> 04:27.090 So writing them out like that in source code can make the numbers clearer. 04:32.780 --> 04:36.890 Literals have a default type, if we don't see what type they are. 04:37.070 --> 04:39.110 A floating point literal will be double. 04:40.820 --> 04:42.590 If we have an integer, it'll be int. 04:42.650 --> 04:44.840 Unless it's too big, in which case it'll be a long. 04:45.380 --> 04:47.570 And if it's too big for long, then it'll be a long long. 04:48.320 --> 04:51.050 And if it's too big for long, long, then you have an overflow. 04:52.930 --> 04:58.510 If you don't want to use these default types, you can add a suffix. So if you want three point 04:58.510 --> 05:01.640 one four one five nine to be a float, then you put an 'f' after it. 05:02.470 --> 05:05.350 That can be upper or lowercase. And there are others. 05:05.530 --> 05:08.650 So, for instance, you can make this number an unsigned long long. 05:11.650 --> 05:18.250 So you can think of this letter 'f' or 'ULL' as a kind of operator, which returns an object of this type. 05:18.730 --> 05:25.330 And in fact, in C++ 11, you can provide your own suffixes which work with your own types and they'll 05:25.360 --> 05:26.590 return an object of that type. 05:29.710 --> 05:31.480 OK, so that's it for this video. 05:31.810 --> 05:34.900 I'll see you next time, but meanwhile, keep coding!