linux/bash: remove text between two chars

Question

I have a file file1 and it contains:

<     16 ./lnx/apps/vlc/tsconf_1.0-11_all.deb
<     16 ./lnx/apps/vlc/vlc
<   2000 ./lnx/apps/vlc/vlc-nox_2.0.8-1_i386.deb
<     16 ./lnx/apps/vlc/vlc-plugin-notify_2.0.8-1_i386.deb
<     32 ./lnx/apps/vlc/vlc-plugin-pulse_2.0.8-1_i386.deb
<     16 ./lnx/cmds/64bit_ubuntu_add_i386
<     16 ./lnx/cmds/acroread_dl
<     16 ./lnx/cmds/dl_from_gdrv
<     16 ./lnx/cmds/dpkg_install_list_in_txt_file
<     16 ./lnx/cmds/find_and_replace
<     16 ./lnx/cmd/pearl_script.ps1
<     16 ./lnx/cmds/rm_using_find
<     16 ./lnx/cmds/wget_dl_whl_ws

I want to remove everything between < and ./ (or between \n and ./) excluding ./, so the output would be the following:

./lnx/apps/vlc/tsconf_1.0-11_all.deb
./lnx/apps/vlc/vlc
./lnx/apps/vlc/vlc-nox_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-notify_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-pulse_2.0.8-1_i386.deb
./lnx/cmds/64bit_ubuntu_add_i386
./lnx/cmds/acroread_dl
./lnx/cmds/dl_from_gdrv
./lnx/cmds/dpkg_install_list_in_txt_file
./lnx/cmds/find_and_replace
./lnx/cmd/pearl_script.ps1
./lnx/cmds/rm_using_find
./lnx/cmds/wget_dl_whl_ws

I tried these commands (their output was the exact input file):

$ sed -n '/</,/ ./ p' file1
$ sed 's/< .* " ."/./' file1
$ sed -e 's/<\n[^ .]>/<\n.>/g' file1
$ sed -e 's/\(<\).*\(.\)/\1\2/' file1

Probably it’s only a simple thing, but I am new at sed/awk/tr/grep/find commands.

Answer 1

$ sed 's/[^.]*//' file
./lnx/apps/vlc/tsconf_1.0-11_all.deb
./lnx/apps/vlc/vlc
./lnx/apps/vlc/vlc-nox_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-notify_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-pulse_2.0.8-1_i386.deb
./lnx/cmds/64bit_ubuntu_add_i386
./lnx/cmds/acroread_dl
./lnx/cmds/dl_from_gdrv
./lnx/cmds/dpkg_install_list_in_txt_file
./lnx/cmds/find_and_replace
./lnx/cmd/pearl_script.ps1
./lnx/cmds/rm_using_find
./lnx/cmds/wget_dl_whl_ws

Answer 2

How about (or am I missing something):

$ awk '$0=$NF' file
./lnx/apps/vlc/tsconf_1.0-11_all.deb
./lnx/apps/vlc/vlc
./lnx/apps/vlc/vlc-nox_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-notify_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-pulse_2.0.8-1_i386.deb
./lnx/cmds/64bit_ubuntu_add_i386
./lnx/cmds/acroread_dl
./lnx/cmds/dl_from_gdrv
./lnx/cmds/dpkg_install_list_in_txt_file
./lnx/cmds/find_and_replace
./lnx/cmd/pearl_script.ps1
./lnx/cmds/rm_using_find
./lnx/cmds/wget_dl_whl_ws

Answer 3

Using sed:

sed -i.bak 's~^<.*[[:space:]]\(\./\)~\1~' file

OR

sed -i.bak 's~^<[^.]*\(\./\)~\1~' file

./lnx/apps/vlc/tsconf_1.0-11_all.deb
./lnx/apps/vlc/vlc
./lnx/apps/vlc/vlc-nox_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-notify_2.0.8-1_i386.deb
./lnx/apps/vlc/vlc-plugin-pulse_2.0.8-1_i386.deb
./lnx/cmds/64bit_ubuntu_add_i386
./lnx/cmds/acroread_dl
./lnx/cmds/dl_from_gdrv
./lnx/cmds/dpkg_install_list_in_txt_file
./lnx/cmds/find_and_replace
./lnx/cmd/pearl_script.ps1
./lnx/cmds/rm_using_find
./lnx/cmds/wget_dl_whl_ws

Assuming there are no directories ending with <space>..

Answer 4

You can use BASH regex matching.

while read -r; do
    [[ $REPLY =~ .*(\.\/.*) ]] && echo ${BASH_REMATCH[1]}
done < file

Answer 5

Bash

while read -r _ _ file_details; do
  echo "$file_details"
done < your_file > your_file.edited

Awk

awk '{ $1=""; $2=""; print; }'

OK, if those first white spaces are really irksome:

awk -F. '{printf ".";print $2}'