Use sed to replace only part of a string

Question

I'm trying to replace bobearl with jim in the following string

"billy" "bobearl" and "johnny"

I can do something like this:

sed 's/bob/jim/' /tmp/text.txt
"billy" "jimearl" and "johnny"

but that leaves earl

I can do this:

sed 's/bob.*/jim/' /tmp/text.txt
"billy" "jim

But that removes everything after the replacement string.

What is the most efficient way to replace bobearl with jim while not changing anything else? I assume my issues has something to do with the quotes but I'm just not sure how to get it to work.

Answer 1

Since bob does not match bobearl your expression s/bob/jim/ will not replace bobearl, it will only replace the bob.

If you match bobearl it will replace bobearl:

$ echo '"billy" "bobearl" and "johnny"' | sed 's/bobearl/jim/'
"billy" "jim" and "johnny"

If you are looking to use a wildcard match without globbing the rest of the line you can do something like this:

 sed -E 's/\bbob\w*?\b/jim/'

• -E use extended regex
• \b matches word boundaries
• \w*? match zero or more word characters non greedy