understanding sed patterns
 

How are patterns in sed handled in this situation?

source: http://gilmation.com/articles/regexp...attern-in-sed/
The following example finds words surrounded by the same number (ab is surrounded by 1 and ef is surrounded by 82) and removes the surrounding number:

The substitute command is s/// as in

So the \1 you have in your formula above is in the 'old' part. The 'new' part has only \2

So if you have a pattern along the lines of 'A B A' then it is replaced with 'B'

Hey, don't delete stuff. Someone searching for a similar problem won't see the key part when this thread turns up in their search results.

Sorry, I didn't know how to handle it :) It was actually much easier than I had thought and I deleted it before you replied, otherwise I wouldn't have deleted it. I'll just rewrite the post then.

Thanks.

I had interpreted the slash between \1 and \2 as the third occurrence, because I hadn't seen \1 in the first part of the substitution before, and I automatically thought that there was a another slash before \1. Thanks.