Splitting and combining files

- [Instructor] Files are really just a bunch of bits. So if we need to, we can slice and dice a file into various parts and put them back together again. Splitting a file might be necessary to store it on a few devices that might not be able to hold the full thing. Or to make individual smaller pieces available for download, or something like that. To split a file, we use the command split. Here in my Pictures folder, I have a picture of a cat. Let's split it into a few pieces and put it back together again.

He looks surprised, but it won't hurt. Here in the terminal, I'll use the command split and then specify -b. This tells split to divide the file by bytes. And I'll tell split to split this file into one megabyte chunks. Then I'll use -d, which appends decimal numbers, instead of strings to the end of each of the parts' file names. Then I'll provide the file to split, cat.jpg. And I'll provide a template for how to name the parts.

In this case, we'll call them cat.jpg.part. And then because of the -d option, the pieces of the file will be called part00, part01, and so on. Now I have four parts of the file and I can see that the first three are each one megabyte and then the remainder is in the last file. This file won't really be useful when it's split like this. Each piece is just a little bucket of bits that don't mean anything without the context of the other parts of the file.

If you're distributing the file or storing it or something like that, you may want to calculate the checksum of the original, and of each part, so you can compare later. Let's calculate the checksum of the original image with md5sum cat.jpg. Okay, we can put the parts back together again with a cat command, either explicitly, or by using Shell shortcuts or scripts. To do that, I'll write cat and then provide the file names of the pieces in the order that I want to put them back together.

Then I'll redirect the output of that to a new file. In this case, I'll call it cat2.jpg. Now let's compare the checksum of the original file with this file. I'll write echo, paste the hash, space cat2.jpg and pipe that to md5sum -c. It looks like everything's okay. I'll switch to my file browser and here's the cat2.jpg file we created.

I'll open it up. We know these pieces have been put back in the right order because we did it explicitly and the checksum matches and the picture looks correct. If we put the pieces back in an incorrect order, we'd have a bad checksum and the resulting data would be incorrect too. Split can create pieces of the file by line number, which can be useful for text, or by size, and there are various other options available as well. I encourage you to explore the man page for split to learn more.