What makes you think that sed is related to bash? It's about as much related as COBOL is...

1 members found this post helpful.

Glad that worked.

As noted in my updated post above, I think the original source of confusion is in use of the '<...>' syntax, which is not common regular expression syntax, although it does seem to be a GNU sed extension with its own quirks. On the other hand, '/^...$/' is basic and universally recognized regular expression syntax.

Learn and stick to common regex or pcre syntax for best results, at least until you have a good feel for their use. Here is a good tutorial to start with, many others available online.

If this question is resolved, please mark it solved with the thread tools at top.

1 members found this post helpful.

No need to escape # in sed, because it has no special meaning in an RE.
The \< and \> are left and right "word boundaries" in many sed versions. But it is not much standardized, so it is not clear if a # is a word character or a non-word character.
Perhaps you can use the line start/end anchors instead
The /g option is pointless (trying multiple substitutions per line).