I have a script file that I am trying to make that will modify the command line parameters withing several start-up scripts. This bash scrip will look for a path and replace the entire line with a new one. Any help would be appreciated, below you will find an example of my code.
Path=/usr/bin/MyApplication
NewArguments= -a1 -b2 -c3
NewCommand="$Path $NewArguments"
sed -i 's,^'"$Path"'*,\'"$NewCommand"',' /root/etc/rc.d/99_start_app.sh
What I'm looking for is converting a line like this:
/usr/bin/MyApplication -x1 -y2 -z3
Into this:
/usr/bin/MyApplication -a1 -b2 -c3 &
Any help with this would be appreciated.
I think you meant this:
sed -i 's,^'"$Path"','"$NewCommand"',' /root/etc/rc.d/99_start_app.sh
...which can be simplified as:
sed -i s,^"$Path","$NewCommand", /root/etc/rc.d/99_start_app.sh
...and probably easier to read as:
sed -i "s,^$Path,$NewCommand," /root/etc/rc.d/99_start_app.sh
Actually the way you assign NewArguments is also incorrect, and I would also improve the sed, so in finally your script could be written like this:
Path=/usr/bin/MyApplication
NewArguments='-a1 -b2 -c3'
sed -i "s,^$Path.*,$Path $NewArguments \&," /root/etc/rc.d/99_start_app.sh
double quotes to expand shell variables
sed "s/$var/$var2/" file
Related
After running command sed -i "/class MySQLPagedResultsajax/i Thing I Wish To Add;" /File/Location/
The result I get are as follows:
Thing I Wish To Add;class MySQLPagedResultsajax
Whereas I want to achieve this:
Thing I Wish To Add;
class MySQLPagedResultsajax
Please Help As I Want A Single Line Command Solution As This Needs To Be Done On Multiple Files And I Will Be Using Excel To Create Separate Commands For Each File.
Try this:
sed -i '/class MySQLPagedResultsajax/s/^/Thing I Wish To Add;\n/'
Or before you are sure, try this first and check the output on console:
sed -e '/class MySQLPagedResultsajax/s/^/Thing I Wish To Add;\n/'
s is the substitute command.
^ means the beginning of the line.
\n a newline.
Just use awk for anything other than a simple s/old/new:
awk '/class MySQLPagedResultsajax/{print "Thing I Wish To Add;"} 1' file
add -i inplace to do pseudo-inplace editing with GNU awk like you can do pseudo-inplace editing with GNU sed by adding -i.
I'm writing a bash script to non-interactively enable hibernation on a linux system. To this end, I need to insert a shell variable that contains slashes on a specific line of a while, inside quotes that are on that line.
The relevant part of the file to b edited looks like this:
GRUB_DEFAULT=0
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR="Manjaro"
GRUB_CMDLINE_LINUX_DEFAULT="quiet"
GRUB_CMDLINE_LINUX=""
I need to change it to this:
GRUB_DEFAULT=0
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR="Manjaro"
GRUB_CMDLINE_LINUX_DEFAULT="quiet resume=/dev/disk/by-partuuid/c5a552c2-fe8f-423a-9037-c35bf090d9c3"
GRUB_CMDLINE_LINUX=""
The added parameter is provided by a shell variable.
I tried this:
sed -i '\*^GRUB_CMDLINE_LINUX_DEFAULT* s*"$* '"$(grub_resume_boot_option)"'"*' /etc/default/grub
Logic that I was aiming for was: "Look for a line that starts with pattern "GRUB_CMDLINE_LINUX_DEFAULT", and replace the last " with the given pattern. Use * as delimiter to preserve the slashes in the expanded variable."
However, the command fails if there are any spaces at the end of the line. Is there any way to make it not take spaces at the end of the line into account?
Also, if there is more simple or readable solution to this, I would be very interested.
You can do:
sed -i 's#^\(GRUB_CMDLINE_LINUX_DEFAULT="quiet\)"$#\1 resume=/dev/disk/by-partuuid/c5a552c2-fe8f-423a-9037-c35bf090d9c3"#' /etc/default/grub
Example:
$ sed 's#^\(GRUB_CMDLINE_LINUX_DEFAULT="quiet\)"$#\1 resume=/dev/disk/by-partuuid/c5a552c2-fe8f-423a-9037-c35bf090d9c3"#' <<<'GRUB_CMDLINE_LINUX_DEFAULT="quiet"'
GRUB_CMDLINE_LINUX_DEFAULT="quiet resume=/dev/disk/by-partuuid/c5a552c2-fe8f-423a-9037-c35bf090d9c3"
From: https://serverfault.com/questions/885684/editing-the-value-of-grub-cmdline-linux-default-thru-bash-script
sed -i 's/^GRUB_CMDLINE_LINUX_DEFAULT="/&newstuff /' /etc/default/grub
This will add "newstuff" to the beginning of the list of params. Examples:
sed 's/^GRUB_CMDLINE_LINUX_DEFAULT="/&newstuff /' <<< 'GRUB_CMDLINE_LINUX_DEFAULT="quiet"'
GRUB_CMDLINE_LINUX_DEFAULT="newstuff quiet"
sed 's/^GRUB_CMDLINE_LINUX_DEFAULT="/&newstuff /' <<< 'GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"'
GRUB_CMDLINE_LINUX_DEFAULT="newstuff quiet splash"
sed -i "s/GRUB_CMDLINE_LINUX_DEFAULT="[^"]*/& $(grub_resume_boot_option)/" /etc/default/grub
I am quite new to shell scripting.
I am scraping a website and the scraped text contains a lot of repetitions. Usually they are the menus on a forum, for example. Mostly, I do this in Python, but I thought that sed command will save me reading and printing the input, loops etc. I want to delete thousands of repeated lines from the same single file. I do not want to copy it to another file, because I will end up with 100 new files. The following is a shadow script which I run from the bash shell.
#!/bin/sed -f
sed -i '/^how$/d' input_file.txt
sed -i '/^is test$/d' input_file.txt
sed -i '/^repeated text/d' input_file.txt
This is the content of the input file:
how to do this task
why it is not working
this is test
Stackoverflow is a very helpful community of programmers
that is test
this is text
repeated text is common
this is repeated text of the above line
Then I run in the shell the following command:
sed -f scriptFile input_file.txt
I get the following error
sed: scriptFile line 2: untermindated `s' command
How can I correct the script, and what is the correct syntax of the command I should use to get it work?
Any help is highly appreciated.
assuming you know what your script is doing, it's very easy to put them into a script. in your case, the script should be:
/^how$/d
/^is test$/d
/^repeated text/d
that's good enough.
to make the script alone to be executable is easy too:
#!/usr/bin/env sed -f
/^how$/d
/^is test$/d
/^repeated text/d
then
chmod +x your_sed_script
./your_sed_script <old >new
here is a very good and compact tutorial. you can learn a lot from it.
following is an example from the site, just in case the link is dead:
If you have a large number of sed commands, you can put them into a file and use
sed -f sedscript <old >new
where sedscript could look like this:
# sed comment - This script changes lower case vowels to upper case
s/a/A/g
s/e/E/g
s/i/I/g
s/o/O/g
s/u/U/g
Wouldn't it be easier to do it with egrep followed by a mv, for example
egrep -v 'pattern1|pattern2|pattern3|...' <input_file.txt >tmpfile.txt
mv tmpfile.txt input_file.txt
Each pattern would describe the lines being deleted, much like in sed. You would not end up with additional files, because the mv removes them.
If you have so many pattern, that you don't want to specify them directly on the command line, you can store them in a file use the -f option of egrep.
I know this is a simple answer and I could probably keep digging around on Google before I stroll across the answer. But I am on a tight schedule and I was hoping for an easy response.
I need to update a variable in ifcfg-eth0 upon an installation. So in other words, this is what needs to happen:
The following variables need to change from:
ONBOOT=no
BOOTPROTO=dhcp
to
ONBOOT=yes
BOOTPROTO=static
Thanks in advance!
Cheers.
sed -i -e '/^ONBOOT=/s|.*|ONBOOT=yes|; /^BOOTPROTO=/s|.*|BOOTPROTO=static|' file
Also try:
sed -i -re 's|^(ONBOOT=).*|\1yes|; s|^(BOOTPROTO=).*|\1static|' file
Or
sed -i -e 's|^\(ONBOOT=\).*|\1yes|; s|^\(BOOTPROTO=\).*|\1static|' file
You can make use of something like this, which lets you define exactly what values you want to add in the file:
$ bootproto="static"
$ sed -r "s/(BOOTPROTO\s*=\s*).*/\1$bootproto/" file
ONBOOT=no
BOOTPROTO=static
And to make the two of them together:
$ onboot="yes"
$ bootproto="static"
$ sed -r -e "s/(ONBOOT\s*=\s*).*/\1$onboot/" -e "s/(BOOTPROTO\s*=\s*).*/\1$bootproto/" file
ONBOOT=yes
BOOTPROTO=static
Explanation
(BOOTPROTO\s*=\s*).* catches a group of text containing BOOTPROTO + any_number_of_spaces + = + any_number_of_spaces. Then, matches the rest of the text, that we want to remove. \1$var prints it back together with the given variable.
\s* keeps the current spaces as they were and then replaces the same line changing the current text with the bash variable $bootproto.
-r is used to catch groups with () instead of \( and \).
To make it edit in place, use sed -i.bak. This will create a backup of the file, file.bak, and the file will be updated with the new content.
All together:
sed -ir -e "s/(ONBOOT\s*=\s*).*/\1$onboot/" -e "s/(BOOTPROTO\s*=\s*).*/\1$bootproto/" file
The similar as the sed solutons with perl
perl -i.bak -pe 's/(ONBOOT)=no/$1=yes/;s/(BOOTPROTO)=dhcp/$1=static/' files....
I have to implement an application in shell programming (Unix/Linux).
I have to search a word from a text file and replace that word with my given word. I have a knowledge on shell and still learning.
I am not expecting source code. Can anybody help me or suggest me or give me some similar solution....
cat abc.txt | grep "pattern" | sed 's/"pattern"/"new pattern"/g'
The above command should work
Thanks,
Regards,
Dheeraj Rampally
Say you are looking for pattern in a file (input.txt) and want to replace it with "new pattern" in another (output.txt)
Here is the main idea, without UUOC:
<input.txt sed 's/"pattern"/"new pattern"/g' >output.txt
todo
Now you need to embed this line in your program. You may want to make it interactive, or a command that you could use with 3 parameters.
edit
I tried to avoid the use of output.txt as a temporary file with this:
<input.txt sed 's/"pattern"/"new pattern"/g' >input.txt
but it empties input.txt for a reason I can't understand. So I tried with a subshell, so:
echo $(<input.txt sed 's/pattern/"new pattern"/g')>input.txt
... but the echo command removes line breaks... still looking.
edit2
From https://unix.stackexchange.com/questions/11067/is-there-a-way-to-modify-a-file-in-place , it looks like writing to the very same file at once it not easy at all. However, I could do what I wanted with sed -i for linux only:
sed -i 's/pattern/"new pattern"/g' input.txt
From sed -i + what the same option in SOLARIS , it looks like there's no alternative, and you must use a temporary file:
sed 's/pattern/"new pattern"/g' input.txt > input.tmp && mv input.tmp input.txt