Help regarding behavior sed regexp query

Shell Programming and Scripting
1

Hi all,

I have one question regarding sed regexp (or any regexp in general),

I have some path like this

C:/Abc/def/ghi/jkl in a file file1

Now if i use following code

cat file1 | sed 's#\(.*\)/.*#\1#'

Now it give me following output
C:/Abc/def/ghi, which is fine

But i just want to know that in this sed code, why it is treating only last forward slash in file1 as the base for creating value of subsection (.*). i mean to say why it is treating C:/Abc/def/ghi as value of subsection. We are having 5 forward slashes in file1, why not it is picking upto C:/Abc/def.

Is that the property of regexp that if a sentence has five forward slashes & i use .*/.*, it will treat forward slash as last forward slash in file.

Can any one please suggest me some article/link to this question. I hope you are getting my question.

Thanks in advance

Sarbjit

2

Hi the regexp in sed is greedy, which means it will try to find the longest possible match.

3

UUOC

sed 's#\(.*\)/.*#\1#' file1

If you want to match only to the first slash, first match only those characters that are not slashes:

sed 's#\([!/]*\)/.*#\1#' file1