WEBVTT 00:00.990 --> 00:07.170 Hello again! In this video, we are going to look at files and streams. You probably have some idea of 00:07.170 --> 00:08.700 what a file means. 00:09.210 --> 00:12.300 C++ has a very minimalistic interpretation. 00:12.780 --> 00:17.310 A file is just a sequence of bytes, which has some file name to identify it. 00:17.940 --> 00:24.180 So we do nt need to know whether the data is stored on a disk or an SD card or some physical device 00:24.180 --> 00:25.350 that is pretending to be a file. 00:26.010 --> 00:30.480 We do not know how exactly it is stored, and we do not care what it is meant to represent. 00:33.530 --> 00:38.060 The C++ library provides fstream objects for interacting with files. 00:38.600 --> 00:43.850 These are very similar to the iostream objects that we have used for interacting with the console, to send 00:44.300 --> 00:46.490 output to the console and read input from it. 00:48.180 --> 00:54.570 These fstream objects always access files sequentially, so that means one byte after another. 00:55.020 --> 00:59.280 You cannot jump forwards or go backwards, and they are always in the same order. 01:01.380 --> 01:07.470 The stream does not know how much data it is going to send or receive, and the stream has no concept of 01:08.100 --> 01:09.270 structure to the data. 01:09.900 --> 01:14.280 So this means, in particular, that fstreams do not understand file formats. 01:17.110 --> 01:20.260 There are four main things you can do with an fstream object. 01:20.590 --> 01:21.430 You can open it. 01:21.760 --> 01:24.400 So this will start interacting with a file. 01:25.390 --> 01:27.970 This will connect to the string object to the file, if you like. 01:28.420 --> 01:31.540 So after that, the file is available for use by your program. 01:32.560 --> 01:33.900 You can read. 01:34.030 --> 01:37.540 So this means that data is copied from the file into your program's memory. 01:38.440 --> 01:38.800 You can 01:38.890 --> 01:39.250 write, 01:39.280 --> 01:44.930 so that means that data is copied from your programs's memory to the file. And you can close the file. 01:45.460 --> 01:49.930 So after you have finished using your file, you can end your interaction session. 01:50.380 --> 01:53.590 And this will close the connection between the stream object and the file. 01:54.010 --> 02:00.070 And then that file is no longer available for use by the program. Unless an fstream object opens it 02:00.130 --> 02:00.400 again. 02:07.280 --> 02:13.250 Each of these operations is performed by calling a function in the operating system's API. The Application 02:13.250 --> 02:14.540 Program Interface. 02:15.200 --> 02:20.570 So the program is going to stop and wait and then the operating system is going to do the read or whatever. 02:21.140 --> 02:28.010 And then when it finishes, the call to the API will return. And the program will then resume executing 02:28.010 --> 02:28.910 your instructions. 02:29.390 --> 02:34.370 So while this call is being performed, your program has to stop and wait. 02:34.610 --> 02:36.620 It cannot execute any of your code. 02:40.790 --> 02:46.560 We cannot use a file until it has been opened. Once we have finished using a file, 02:46.580 --> 02:47.540 we should close it. 02:48.620 --> 02:53.420 If you are writing to a file, this will make sure that all the data that you wrote is actually saved 02:53.720 --> 02:54.230 correctly. 02:55.070 --> 03:01.070 And also, operating systems have a limit on the amount of files this program can have open at any one 03:01.070 --> 03:01.490 time. 03:02.210 --> 03:05.540 If you try to open too many files, the operating system will stop you. 03:05.930 --> 03:07.820 So this is a bit like allocating memory. 03:08.210 --> 03:13.070 If your program allocates too much memory, the operating system will stop you from allocating more 03:13.070 --> 03:16.670 memory. I actually fell into this once, 03:16.670 --> 03:20.360 many years ago. I had a loop that opened to file, but did not close it. 03:21.440 --> 03:25.790 And then eventually, after 40 iterations or whatever it was, the operating system said, 03:26.180 --> 03:26.990 "Sorry, no more!" 03:29.990 --> 03:35.480 When you compile a C++ program into a binary, the compiler will add some extra runtime code to 03:35.480 --> 03:35.990 the program. 03:36.380 --> 03:38.930 So this will set the program up and then call main(). 03:39.440 --> 03:43.940 And then when main() returns, this code will do any clearing up that is needed. 03:44.300 --> 03:48.740 And one of the things it does is that it will go through all the files that the program has opened 03:49.220 --> 03:50.450 and make sure they are closed. 03:51.800 --> 03:55.180 So in theory, you do not actually need to close any files. 03:55.190 --> 03:59.300 But, as with releasing memory, it is a good idea to do it as soon as you can. 03:59.780 --> 04:05.870 It is actually more important with files, because other programs can access files and - well, you can have 04:05.870 --> 04:09.620 file looking, but C++ does not support that directly. 04:10.160 --> 04:16.070 So you could have other programs trying to use your file. So you really want to have your file 04:16.070 --> 04:17.870 open for the shortest possible time. 04:23.800 --> 04:26.830 And then one final point before we go on. 04:27.460 --> 04:34.210 When you send or receive data with a file, it is not going to be actually sent one byte at a time. 04:35.290 --> 04:39.910 The reason is that interacting with physical devices is very time-consuming. 04:40.420 --> 04:46.120 You know, it takes thousands of processor instructions. So you really want to minimize the number 04:46.120 --> 04:47.110 of actual transfers. 04:48.520 --> 04:52.030 So what happens is that the data is gathered up into a memory buffer. 04:52.360 --> 04:56.680 And then when it reaches the optimal size, the data gets sent off to the device. 04:59.470 --> 05:04.210 So this reduces the number of actual transfers, so that reduces the number of times your program has 05:04.210 --> 05:04.510 to stop and wait. 05:05.380 --> 05:11.400 However, there is a disadvantage to this. If your program waits too long between transferring data, 05:11.860 --> 05:18.850 then the discrepancy, between what the file says, and what your program thinks the file says, will get bigger 05:18.850 --> 05:19.400 and bigger. 05:19.900 --> 05:21.580 And that can cause problems. 05:21.910 --> 05:22.840 But we will come back to that. 05:23.470 --> 05:25.150 Okay, so that's it for this video. 05:25.480 --> 05:26.290 I'll see you next time. 05:26.290 --> 05:28.510 But meanwhile, keep coding!