[MUSIC] Hi, and welcome to lesson eight, file input/output. This week, we'll be talking about how to read data from and write data to files. So what is a file? Well, a file is an area in permanent storage, typically a disk drive, that stores information. Files are managed by the operating system of your computer. They can be copied, moved, renamed, deleted, et cetera, and they can be accessed by programs running on your computer. A very important use of files is to exchange data between different programs and different computers. In that case, a file is input to one program, and is output from a different program. We abbreviate the interchange of information between programs and files as file I/O for file input and output. For example, you might want to write a document with a word processor and generate a PDF. And you can give that PDF file to someone else who can open it with Adobe Acrobat on another computer and print it. And files can be used to exchange data between the physical world and computers. For example, an EKG machine, which measures your heart activity, will save the data into a file. And your doctor can open that file on a computer and evaluate your health. The most common use of files is to save your work so you can continue with it later. Examples of that include, documents that you write on word processors or spreadsheets. And extends to the files we've been seeing MATLAB save when he needed to save data or functions like MATLAB files and M-files. But you can handle much more than Mat-files and M-files with MATLAB. As we'll soon see, with MATLAB, you can read and write text files, binary files, and even Microsoft Excel spreadsheets. At the beginning of our journey, we saw how to use the user interface of the MATLAB desktop to change folders, and how MATLAB shows the contents of the current folder in this window over here. Sometimes however, we may need to have our MATLAB program itself change the current folder, or get a list of the names of the files in there. Fortunately, MATLAB provides commands to do just that. To get the fully qualified name of the current folder, we can use the PWD command. PWD stands for print working directory. It originates from the UNIX operating system, where folders are called directories. And the command returns a string, containing the complete path of the current folder, complete path or fully qualified name. It's exactly the same path as that shown up here in the address line. We haven't talked about the address line since lesson two, so you may not have thought about it since then. But it shows the current folder, which is called Lesson 8, and shows that it's contained in the folder, Tutorials, which is contained in the folder, Computer Programming with MATLAB, et cetera. More commonly, we say that Lesson Eight is a child of Tutorials, which is a child of Computer Programming with MATLAB, and so forth. Or to put it the other way, computer programming with MATLAB is the parent of Tutorials, which is the parent of Lesson 08. In operating systems, they're all single parent families. The only difference between here and here, is that these little arrows in the address line are represented by slashes in the output from PWD. We're doing this on a Mac, but if you're using Windows, the output from PWD will be slightly different. These slashes turn into backslashes, and there'll be a letter and a colon at the front representing the hard drive. So if we're using Windows, the output would look like this. To make this Windows like string here, I've used the built-in function strrep, which means string replace. It replaces every slash that's in the output from pwd here with the backslash. You can see how the Syntax and semantics of this command work, but you should look it up in the help. It's a handy function. To get a list of the files and subfolders in the current folder, you can use the command LS, which stands for list. Use an output argument. LS produces an array of chars. Let's do that. You can see here that it's a 1 by 198 string of chars. To change folders, MATLAB provides the CD command. This cd doesn't mean compact disc or certificate of deposit. It stands for change directory. You have to give it a string as an argument to tell it the directory you want to go to. Suppose your program's organization is such that it starts working in the current folder, where ever that folder may be, then it needs to move to the parent. Like here, for example, if it started in lesson 08, it would want to move to the parent, Tutorials. You're always going to start to program in lesson eight, or some other child of tutorials, like lesson seven. Then you could just store the complete path name of tutorials in a string in a program like this. And then, when it was time to change to the parent, you could do this. Before I hit return, let's look up here and watch as I hit return. So, here comes return. And you see, that we're now up in tutorials. Let's go back to lesson 08. There, we're back in lesson 08. But suppose your program was run from another directory. Then the parent that it needs to move to will be different, and so using this complete path name idea just won't work. This is a very common situation. And for this situation, our program can use a so-called relative path name. What that means is that, instead of specifying the complete path name, we can specify the target folder relative to the current folder. Or to press the family metaphor a little further, we specify relative of the current folder. In this case, the relative is the parent, and there's a special syntax for specifying the parent. We want to go to the parent of the current folder, we can do it like this. Two dots. Now we're back at the tutorials again. The two dots means parent folder. Now let's use LS to see what's here in tutorials. Of course, we can see it but, the program couldn't. And if your program needs to see what's inside of one of the children of tutorials, it can give its name as an argument to LS, without the complete path, like this. Not much in there. By the way, this is how the function would do it, it would give the string containing the target directory in parentheses. But if we want to do that ourselves in the command window, as in fact we just did, there's a second way to do it. Like this, does the same thing. When function takes a string as an argument, the parentheses can be omitted. When there are no spaces in the name, the quotes can be omitted too, and this is whether you write the function or it's a built in function. That's how commands like help work. The command help square root, for example, is different in form. In other words, different in syntax from this command. But both have the same meaning in, in other words, they both have the same semantics. Now let's use cd to go back to lesson eight, using relative naming. By the way, the dot dot, which means parent, which we saw a little earlier, can be used multiple times within the path. Before we show you that, you may have noticed that I slipped a subfolder into Lesson 8 a few minutes ago. It's called Additional Examples. Let's go there. I cheated and just clicked. That's just one file in there called knothole, it's an M-file with a function in it that checks to see if its argument is not a whole number. Anyway, now that we're way down here, in additional examples, we can leap back up to tutorials, which is the grandparent like this. There. Back to tutorials. Here's some other commands that allow you to do things with folders and files. This one makes a new directory. It just showed up over here. And this one removes the directory. It just disappeared over here. But it removes only empty directories. It'll do nothing, except give an error if you try to remove a folder with anything in it. Let's go into eight, and try to remove my additional examples folder, which you just saw has a file in it. I'd hate to lose that. Oops. Then as you can see, it's still there. And there are three commands for doing things to files, and the names of the files tell what they do. They're called move file, copy file and delete. Each of these commands is relatively simple to use. As you can see by reading their descriptions with the help command, but I'll give you a warning about delete. Delete gives you no warning. When you tell it to delete a file, no pop-up will show up saying, are you sure and that sort of thing. But all is not lost. There's a setting in preferences up here, that will cause the file to be sent to trash. It's probably a good idea to set that. Okay, I've started working on another project, since I did all that stuff with folders and files. I want to show you all the variables I've got in the workspace now. There, isn't that nice? In week one, we mentioned briefly how you can save your current workspace with the save command, like this. And now if you look over here in the current folder, you will see this file appear matlab.mat. It's got everything that we had in our workspace saved in it. Then you can safely quit MATLAB and later when you are ready to continue your work, you can start MATLAB and load the workspace back with this load command here. And when you check your workspace, it's all the same. Course I didn't actually quit MATLAB just now, so you might not be convinced. Maybe this will convince you. There, everything's gone. And everything's back. The only thing you have to be careful about, is that if you have an existing variable in the workspace with the same name as the one you're loading, the load operation will overwrite the existing value. Usually that isn't a problem though, if the last thing you do is save when you quit and the first thing you do is load when you start again. By the way, you should look up the function startup, which is executed when MATLAB starts, and finish, which is executed when MATLAB quits. You can put a load command in startup, and a save command in finish. Sometimes that's all you need to preserve your variables. MATLAB provides even more flexibility in loading and saving variables. You don't have to save the entire workspace. Instead you can list the variables you want to save, and similarly you can load only a subset of the variables from a file, using their names. Finally, you don't have to use the default file name matlab.mat, like right here. You can supply your own favorite file name. Check this out. Here I've specified that I want to save variables into a file called my_data_file.mat. You can see it's showed up right over here. And then I specify the data I want to save into it. Data s and a. Now let's clear. Run whos. There's nothing there. And now let's just load my data file. It's back. What if I just wanted to load, say, S and A? Let's try that. All I get is S and A. These load and save commands supply a flexible means to provide data persistence for your MATLAB programs, from one running of MATLAB to the next. [MUSIC] [APPLAUSE]
