sed and awk

A stream editor that reads a text line by line, applies commands to each one (substitute, delete, print) and writes the result to its output.

It does not modify the file it reads: without -i, the result goes to standard output. The most common command is s/pattern/replacement/.

a-b+c

The number at the end of s names the occurrence to replace: the second dash becomes +, the first stays. With no number it is the first; with g, all of them.

In sed, the s/a/b/ command replaces only the first occurrence on each line; to replace them all, add the g flag.

echo a-b-c | sed 's/-/+/' gives a+b-c and with g a+b+c. A number instead, as in s/-/+/2, targets the second occurrence: a-b+c.

A text-processing language that splits each line into fields and runs an action on those that match a pattern.

A program is a series of pattern { action }. With no pattern, the action applies to every line; with no action, the line is printed.

sed 's#usr/sbin#bin#' paths.txt

The path's slashes are read as s separators: sed finds too many parts and takes bin/ for a flag. Change the separator, here #, or protect each slash with \. g and -n change nothing.