sed regular expression 1

[ranadhir]

We have a file with entries like
good/bad/ugly.
We are able to retrieve the last bit(ugly) with
sed -e 's/.*\///g' testfilelist.txt
But we wish to achieve the same by creating 'data regions'.
We are trying

sed -e 's/(.*/)([^/]\+)$/\2/g' testfilelist.txt
But we are getting
sed: -e expression #1, char 12: Unknown option to `s'.
How do we print out the first ,second and third entries in the line above by using this approach?

 

[feherke]

Hi

Usually you have to escape the parethesis ( () ) in [tt]sed[/tt]. But no idea what version you use.

Feherke.

http://rootshell.be/~feherke/

 

[feherke]

Hi

By the way, looks like would be easier to use [tt]awk[/tt].

Feherke.

http://rootshell.be/~feherke/

 

[feherke]

Hi

Oops, you have to escape the slashes ( / ) too. To have more clear expression, use other character as delimiter. I prefer comma ( , ) :

sed -e 's,\(.*/\)\([^/]\+\)$,\2,g' /input/file

Feherke.

http://rootshell.be/~feherke/

 

[ranadhir]

tried this
sed -e 's/\(.*/)\([^/]\+)$\/\2/g' testfilelist.txt
But that does not improve matters.
Am using GNU sed version 4.0.7
Is this a proper attempt at partitioning into data regions?

 

[PHV]

awk -F/ '{print $2}' testfilelist.txt

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[ranadhir]

Solved!!
The final form looks like quite scary - but it works
sed -e 's/\(.*\/\)\([^\/]\+\)$/\2/g' testfilelist.txt
Awk is definitely much more simpler;but this was just an attempt to understand the data region concept ,to use in perl

 

[feherke]

Hi

this was just an attempt to understand the data region concept ,to use in perl

There is no "data region" mentioned in [tt]man perlre[/tt]. If you refer to the things between parenthesis ( () ), those are called groups.

Feherke.

http://rootshell.be/~feherke/

 

[PHV]

to use in perl
In perl, why not simply use the split function ?

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[ranadhir]

What i meant is the following construct in perl:
next unless($line=~m/([^\/]+)\/([^\/]+)\/([^\/]+)$/);
This too works fine - the only difference that i see is to not use the escape characters.

 

[feherke]

Hi

Usually you have to escape the parethesis ( () ) in sed.

the only difference that i see is to not use the escape characters.

That is a difference between the regular expression syntax of the two programs.

You could try txt2regex, a tough regular expression editor which can display multiple syntaxes.

http://txt2regex.sourceforge.net/

Some [tt]sed[/tt] versions, like GNU [tt]sed[/tt] has a -r switch to use the typical syntax :

-r, --regexp-extended
use extended regular expressions in the script.

Feherke.

http://rootshell.be/~feherke/