Using sed

sed is one of the most powerful commands when it comes to data wrangling in the shell. It’s called a “stream editor” because you can use it to make edits to a “stream” of text, which could either be a file or the output of another program.

Find-and-Replace

The simplest way to use sed is for finding a word and replacing it with another word:

cat input.txt | sed 's/hello/hi/g'

Let’s look at this in detail:

• sed takes as its argument a “command string”, which tells it what to do
• The first letter of the command string is the command itself. In this case we used s, the “substitute” command. This is the general find-and-replace command, and is likely the most common one you’ll use.
• The next symbol in the string is a “divider”; in this case we used a forward-slash /. This divider is used to separate the different parts of the command string.
• The second part of the command string is the “target”, i.e., the string to be replaced. In this case we are replacing the word hello.
• The third part of the command string is the “replacement”, i.e., the string to insert. In this case we are replacing hello with hi.
• The final part of the command string is a “flag”. We could leave this blank, but in this case we used the g (global) flag. This tells sed to replace every instance of hello, not just the first one on each line.

Regular Expressions

We can also use sed with regular expressions, just like we did with grep in our previous post.