UNIX for Dummies Questions & Answers

Link explains various characters and their usage with sed:

[Chapter 3] 3.2 A Line-Up of Characters

* does not work by itself in a regular expression. It means 'zero or more of the previous character'.

. means 'any character'.

So if you wanted a, something, c, you could do a.*c

Thanks in2nix4life and Corona688,

I am gonna spend some more time to get a better picture, thanks a lot for your help.

It is probably a problem in understanding, not in using:

Suppose you have a string

Code:

abc

and you want to match it, based on the "a" and the "c" occurring. You are looking for one occurrence of "a", followed by whatever, followed by one occurrence of "c". But if you look for "zero (or more) occurrences of 'a'", you basically don't care for the "a" at all - "zero or more" can also be "zero", which means there is no "a" there.

This begs the question what an expression like "a*" is good for, if it means that the "a" is optional at all. Well, suppose the following:

Code:

/xa*y/

This would match any "x", followed by zero or more "a"s, followed by "y". Strings matched by this would be:

Code:

xy
xay
xaaaaay
..etc.

Not matched would be:

Code:

xby
xbcdefy
xblablay
..etc.

In other words: our expression means that between "x" and "y" only "a"s are allowed - regardless how many, even zero - but if there is anything in between "x" and "y" it has to be "a"s and "a"s alone.

I hope this helps.

bakunin

Thanks Bakunin,

That was really helpful......

Thanks again....