Related
How would I use sed to delete all lines in a text file that contain a specific string?
To remove the line and print the output to standard out:
sed '/pattern to match/d' ./infile
To directly modify the file – does not work with BSD sed:
sed -i '/pattern to match/d' ./infile
Same, but for BSD sed (Mac OS X and FreeBSD) – does not work with GNU sed:
sed -i '' '/pattern to match/d' ./infile
To directly modify the file (and create a backup) – works with BSD and GNU sed:
sed -i.bak '/pattern to match/d' ./infile
There are many other ways to delete lines with specific string besides sed:
AWK
awk '!/pattern/' file > temp && mv temp file
Ruby (1.9+)
ruby -i.bak -ne 'print if not /test/' file
Perl
perl -ni.bak -e "print unless /pattern/" file
Shell (bash 3.2 and later)
while read -r line
do
[[ ! $line =~ pattern ]] && echo "$line"
done <file > o
mv o file
GNU grep
grep -v "pattern" file > temp && mv temp file
And of course sed (printing the inverse is faster than actual deletion):
sed -n '/pattern/!p' file
You can use sed to replace lines in place in a file. However, it seems to be much slower than using grep for the inverse into a second file and then moving the second file over the original.
e.g.
sed -i '/pattern/d' filename
or
grep -v "pattern" filename > filename2; mv filename2 filename
The first command takes 3 times longer on my machine anyway.
The easy way to do it, with GNU sed:
sed --in-place '/some string here/d' yourfile
You may consider using ex (which is a standard Unix command-based editor):
ex +g/match/d -cwq file
where:
+ executes given Ex command (man ex), same as -c which executes wq (write and quit)
g/match/d - Ex command to delete lines with given match, see: Power of g
The above example is a POSIX-compliant method for in-place editing a file as per this post at Unix.SE and POSIX specifications for ex.
The difference with sed is that:
sed is a Stream EDitor, not a file editor.BashFAQ
Unless you enjoy unportable code, I/O overhead and some other bad side effects. So basically some parameters (such as in-place/-i) are non-standard FreeBSD extensions and may not be available on other operating systems.
I was struggling with this on Mac. Plus, I needed to do it using variable replacement.
So I used:
sed -i '' "/$pattern/d" $file
where $file is the file where deletion is needed and $pattern is the pattern to be matched for deletion.
I picked the '' from this comment.
The thing to note here is use of double quotes in "/$pattern/d". Variable won't work when we use single quotes.
You can also use this:
grep -v 'pattern' filename
Here -v will print only other than your pattern (that means invert match).
To get a inplace like result with grep you can do this:
echo "$(grep -v "pattern" filename)" >filename
I have made a small benchmark with a file which contains approximately 345 000 lines. The way with grep seems to be around 15 times faster than the sed method in this case.
I have tried both with and without the setting LC_ALL=C, it does not seem change the timings significantly. The search string (CDGA_00004.pdbqt.gz.tar) is somewhere in the middle of the file.
Here are the commands and the timings:
time sed -i "/CDGA_00004.pdbqt.gz.tar/d" /tmp/input.txt
real 0m0.711s
user 0m0.179s
sys 0m0.530s
time perl -ni -e 'print unless /CDGA_00004.pdbqt.gz.tar/' /tmp/input.txt
real 0m0.105s
user 0m0.088s
sys 0m0.016s
time (grep -v CDGA_00004.pdbqt.gz.tar /tmp/input.txt > /tmp/input.tmp; mv /tmp/input.tmp /tmp/input.txt )
real 0m0.046s
user 0m0.014s
sys 0m0.019s
Delete lines from all files that match the match
grep -rl 'text_to_search' . | xargs sed -i '/text_to_search/d'
SED:
'/James\|John/d'
-n '/James\|John/!p'
AWK:
'!/James|John/'
/James|John/ {next;} {print}
GREP:
-v 'James\|John'
perl -i -nle'/regexp/||print' file1 file2 file3
perl -i.bk -nle'/regexp/||print' file1 file2 file3
The first command edits the file(s) inplace (-i).
The second command does the same thing but keeps a copy or backup of the original file(s) by adding .bk to the file names (.bk can be changed to anything).
You can also delete a range of lines in a file.
For example to delete stored procedures in a SQL file.
sed '/CREATE PROCEDURE.*/,/END ;/d' sqllines.sql
This will remove all lines between CREATE PROCEDURE and END ;.
I have cleaned up many sql files withe this sed command.
echo -e "/thing_to_delete\ndd\033:x\n" | vim file_to_edit.txt
Just in case someone wants to do it for exact matches of strings, you can use the -w flag in grep - w for whole. That is, for example if you want to delete the lines that have number 11, but keep the lines with number 111:
-bash-4.1$ head file
1
11
111
-bash-4.1$ grep -v "11" file
1
-bash-4.1$ grep -w -v "11" file
1
111
It also works with the -f flag if you want to exclude several exact patterns at once. If "blacklist" is a file with several patterns on each line that you want to delete from "file":
grep -w -v -f blacklist file
to show the treated text in console
cat filename | sed '/text to remove/d'
to save treated text into a file
cat filename | sed '/text to remove/d' > newfile
to append treated text info an existing file
cat filename | sed '/text to remove/d' >> newfile
to treat already treated text, in this case remove more lines of what has been removed
cat filename | sed '/text to remove/d' | sed '/remove this too/d' | more
the | more will show text in chunks of one page at a time.
Curiously enough, the accepted answer does not actually answer the question directly. The question asks about using sed to replace a string, but the answer seems to presuppose knowledge of how to convert an arbitrary string into a regex.
Many programming language libraries have a function to perform such a transformation, e.g.
python: re.escape(STRING)
ruby: Regexp.escape(STRING)
java: Pattern.quote(STRING)
But how to do it on the command line?
Since this is a sed-oriented question, one approach would be to use sed itself:
sed 's/\([\[/({.*+^$?]\)/\\\1/g'
So given an arbitrary string $STRING we could write something like:
re=$(sed 's/\([\[({.*+^$?]\)/\\\1/g' <<< "$STRING")
sed "/$re/d" FILE
or as a one-liner:
sed "/$(sed 's/\([\[/({.*+^$?]\)/\\\1/g' <<< "$STRING")/d"
with variations as described elsewhere on this page.
cat filename | grep -v "pattern" > filename.1
mv filename.1 filename
You can use good old ed to edit a file in a similar fashion to the answer that uses ex. The big difference in this case is that ed takes its commands via standard input, not as command line arguments like ex can. When using it in a script, the usual way to accomodate this is to use printf to pipe commands to it:
printf "%s\n" "g/pattern/d" w | ed -s filename
or with a heredoc:
ed -s filename <<EOF
g/pattern/d
w
EOF
This solution is for doing the same operation on multiple file.
for file in *.txt; do grep -v "Matching Text" $file > temp_file.txt; mv temp_file.txt $file; done
I found most of the answers not useful for me, If you use vim I found this very easy and straightforward:
:g/<pattern>/d
Source
I've writted a sed script to replace all ^^ with NULL. It seems though that sed is only catching a pair, but not including the second in that pair as it continues to search.
echo "^^^^" | sed 's/\^\^/\^NULL\^/g'
produces
^NULL^^NULL^
when it should produce
^NULL^NULL^NULL^
Try with a loop to apply your command again to modified pattern space:
echo "^^^^" | sed ':a;s/\^\^/\^NULL\^/;t a;'
To edit a file in place on OSX, try the -i flag and multiline command:
sed -i '' ':a
s/\^\^/\^NULL\^/
t a' file
With GNU sed:
sed -i ':a;s/\^\^/\^NULL\^/;t a;' file
or simply redirect the command to a temporary file before renaming it:
sed ':a;s/\^\^/\^NULL\^/;t a;' file > tmp && mv tmp file
I really like SLePort solution, but since it is not working for you, you can try with (tested on Linux, not Mac):
echo "^^^^" | sed 's/\^\^/\^NULL\^/g; s//\^NULL\^/g'
It is doing the same as the former solution, but explicitly, not looping with tags.
You can omit the pattern in the second command and sed will use the previous pattern.
So, disclaimer: I am pretty new to using bash and zsh, so there is a chance the answer is really simple. Nonetheless. I checked previous postings and couldn't find anything. (edit: I have tried this in both bash and zsh shells- same problem.)
I have a directory with many files and am trying to remove the first line from each file.
So say the directory contains: file1.txt file2.txt file3.txt ... etc.
I am using the sed command (non-GNU):
sed -i -e "1d" *.txt
For some reason, this is only removing the first line of the first file. I thought that the *.txt would affect all files matching the pattern in directory. Strangely, it is creating the file duplicates with -e appended, but both the duplicate and original are the same.
I tried this with other commands (e.g. ls *.txt) and it works fine. Is there something about sed I am missing?
Thank you in advance.
Different versions of sed in differing operating systems support various parameters.
OpenBSD (5.4) sed
The -i flag is unavailable. You can use the following /bin/sh syntax:
for i in *.txt
do
f=`mktemp -p .`
sed -e "1d" "${i}" > "${f}" && mv -- "${f}" "${i}"
done
FreeBSD (11-CURRENT) sed
The -i flag requires an extension, even if it's empty. Thus must be written as sed -i "" -e "1d" *.txt
GNU sed
This looks to see if the argument following -i is another option (or possibly a command). If so, it assumes an in-place modification. If it appears to be a file extension such as ".bak", it will rename the original with the ".bak" and then modify it into the original file's name.
There might be other variations on other platforms, but those are the three I have at hand.
use it without -e !
for one file use:
sed -i '1d' filename
for all files use :
sed -i '1d' *.txt
or
files=/path/to/files/*.extension ; for var in $files ; do sed -i '1d' $var ; done
.for me i use ubuntu and debian based systems , this method is working for me 100% , but for other platformes i'm not sure , so this is other method :
replace first line with emty pattern , and remove empty lines , (double commands):
for files in $(ls /path/to/files/*.txt); do sed -i "s/$(head -1 "$files")//g" "$files" ; sed -i '/^$/d' "$files" ; done
Note: if your files contain splash '/' , then it will give error , so in this case sed command should look like this ( sed -i "s[$(head -1 "$files")[[g" )
hope that's what you're looking for :)
The issue here is that the line number isn't reset when sed opens a new file, so 1 only matches the first line of the first file.
One solution is to use a shell loop, calling sed once for each file. Gumnos' answer shows how to do this in the most widely compatible way, although if you have a version of sed supporting the -i flag, you could do this instead:
for i in *.txt; do
sed -i.bak '1d' "$i"
done
It is possible to avoid creating the backup file by passing an empty suffix but personally, I don't think it's such a bad thing. One day you'll be grateful for it!
It appears that you're not working with GNU tools but if you were, I would recommend using GNU awk for this task. The variable FNR is useful here, as it keeps track of the record number for each file individually, allowing you to do this:
gawk -i inplace 'FNR>1' *.txt
Using the inplace extension, this allows you to remove the first line from each of your files, by only printing the lines where FNR is greater than 1.
Testing it out:
$ seq 5 > file1
$ seq 5 > file2
$ gawk -i inplace 'FNR>1' file1 file2
$ cat file1
2
3
4
5
$ cat file2
2
3
4
5
The last argument you are passing to the Sed is the problem
try something like this.
var=(`find *txt`)
for file in "${var[#]}"
do
sed -i -e 1d $file
done
This did the trick for me.
I need to add the following line to the end of a config file:
include "/configs/projectname.conf"
to a file called lighttpd.conf
I am looking into using sed to do this, but I can't work out how.
How would I only insert it if the line doesn't already exist?
Just keep it simple :)
grep + echo should suffice:
grep -qxF 'include "/configs/projectname.conf"' foo.bar || echo 'include "/configs/projectname.conf"' >> foo.bar
-q be quiet
-x match the whole line
-F pattern is a plain string
https://linux.die.net/man/1/grep
Edit:
incorporated #cerin and #thijs-wouters suggestions.
This would be a clean, readable and reusable solution using grep and echo to add a line to a file only if it doesn't already exist:
LINE='include "/configs/projectname.conf"'
FILE='lighttpd.conf'
grep -qF -- "$LINE" "$FILE" || echo "$LINE" >> "$FILE"
If you need to match the whole line use grep -xqF
Add -s to ignore errors when the file does not exist, creating a new file with just that line.
Try this:
grep -q '^option' file && sed -i 's/^option.*/option=value/' file || echo 'option=value' >> file
Using sed, the simplest syntax:
sed \
-e '/^\(option=\).*/{s//\1value/;:a;n;ba;q}' \
-e '$aoption=value' filename
This would replace the parameter if it exists, else would add it to the bottom of the file.
Use the -i option if you want to edit the file in-place.
If you want to accept and keep white spaces, and in addition to remove the comment, if the line already exists, but is commented out, write:
sed -i \
-e '/^#\?\(\s*option\s*=\s*\).*/{s//\1value/;:a;n;ba;q}' \
-e '$aoption=value' filename
Please note that neither option nor value must contain a slash /, or you will have to escape it to \/.
To use bash-variables $option and $value, you could write:
sed -i \
-e '/^#\?\(\s*'${option//\//\\/}'\s*=\s*\).*/{s//\1'${value//\//\\/}'/;:a;n;ba;q}' \
-e '$a'${option//\//\\/}'='${value//\//\\/} filename
The bash expression ${option//\//\\/} quotes slashes, it replaces all / with \/.
Note: Just trapped into a problem. In bash you may quote "${option//\//\\/}", but in the sh of busybox, this does not work, so you should avoid the quotes, at least in non-bourne-shells.
All combined in a bash function:
# call option with parameters: $1=name $2=value $3=file
function option() {
name=${1//\//\\/}
value=${2//\//\\/}
sed -i \
-e '/^#\?\(\s*'"${name}"'\s*=\s*\).*/{s//\1'"${value}"'/;:a;n;ba;q}' \
-e '$a'"${name}"'='"${value}" $3
}
Explanation:
/^\(option=\).*/: Match lines that start with option= and (.*) ignore everything after the =. The \(…\) encloses the part we will reuse as \1later.
/^#?(\s*'"${option//////}"'\s*=\s*).*/: Ignore commented out code with # at the begin of line. \? means «optional». The comment will be removed, because it is outside of the copied part in \(…\). \s* means «any number of white spaces» (space, tabulator). White spaces are copied, since they are within \(…\), so you do not lose formatting.
/^\(option=\).*/{…}: If matches a line /…/, then execute the next command. Command to execute is not a single command, but a block {…}.
s//…/: Search and replace. Since the search term is empty //, it applies to the last match, which was /^\(option=\).*/.
s//\1value/: Replace the last match with everything in (…), referenced by \1and the textvalue`
:a;n;ba;q: Set label a, then read next line n, then branch b (or goto) back to label a, that means: read all lines up to the end of file, so after the first match, just fetch all following lines without further processing. Then q quit and therefore ignore everything else.
$aoption=value: At the end of file $, append a the text option=value
More information on sed and a command overview is on my blog:
https://marc.wäckerlin.ch/computer/stream-editor-sed-overview-and-reference
If writing to a protected file, #drAlberT and #rubo77 's answers might not work for you since one can't sudo >>. A similarly simple solution, then, would be to use tee --append (or, on MacOS, tee -a):
LINE='include "/configs/projectname.conf"'
FILE=lighttpd.conf
grep -qF "$LINE" "$FILE" || echo "$LINE" | sudo tee --append "$FILE"
Here's a sed version:
sed -e '\|include "/configs/projectname.conf"|h; ${x;s/incl//;{g;t};a\' -e 'include "/configs/projectname.conf"' -e '}' file
If your string is in a variable:
string='include "/configs/projectname.conf"'
sed -e "\|$string|h; \${x;s|$string||;{g;t};a\\" -e "$string" -e "}" file
If, one day, someone else have to deal with this code as "legacy code", then that person will be grateful if you write a less exoteric code, such as
grep -q -F 'include "/configs/projectname.conf"' lighttpd.conf
if [ $? -ne 0 ]; then
echo 'include "/configs/projectname.conf"' >> lighttpd.conf
fi
another sed solution is to always append it on the last line and delete a pre existing one.
sed -e '$a\' -e '<your-entry>' -e "/<your-entry-properly-escaped>/d"
"properly-escaped" means to put a regex that matches your entry, i.e. to escape all regex controls from your actual entry, i.e. to put a backslash in front of ^$/*?+().
this might fail on the last line of your file or if there's no dangling newline, I'm not sure, but that could be dealt with by some nifty branching...
Here is a one-liner sed which does the job inline. Note that it preserves the location of the variable and its indentation in the file when it exists. This is often important for the context, like when there are comments around or when the variable is in an indented block. Any solution based on "delete-then-append" paradigm fails badly at this.
sed -i '/^[ \t]*option=/{h;s/=.*/=value/};${x;/^$/{s//option=value/;H};x}' test.conf
With a generic pair of variable/value you can write it this way:
var=c
val='12 34' # it handles spaces nicely btw
sed -i '/^[ \t]*'"$var"'=/{h;s/=.*/='"$val"'/};${x;/^$/{s//c='"$val"'/;H};x}' test.conf
Finally, if you want also to keep inline comments, you can do it with a catch group. E.g. if test.conf contains the following:
a=123
# Here is "c":
c=999 # with its own comment and indent
b=234
d=567
Then running this
var='c'
val='"yay"'
sed -i '/^[ \t]*'"$var"'=/{h;s/=[^#]*\(.*\)/='"$val"'\1/;s/'"$val"'#/'"$val"' #/};${x;/^$/{s//'"$var"'='"$val"'/;H};x}' test.conf
Produces that:
a=123
# Here is "c":
c="yay" # with its own comment and indent
b=234
d=567
As an awk-only one-liner:
awk -v s=option=value '/^option=/{$0=s;f=1} {a[++n]=$0} END{if(!f)a[++n]=s;for(i=1;i<=n;i++)print a[i]>ARGV[1]}' file
ARGV[1] is your input file. It is opened and written to in the for loop of theEND block. Opening file for output in the END block replaces the need for utilities like sponge or writing to a temporary file and then mving the temporary file to file.
The two assignments to array a[] accumulate all output lines into a. if(!f)a[++n]=s appends the new option=value if the main awk loop couldn't find option in file.
I have added some spaces (not many) for readability, but you really need just one space in the whole awk program, the space after print.
If file includes # comments they will be preserved.
Here's an awk implementation
/^option *=/ {
print "option=value"; # print this instead of the original line
done=1; # set a flag, that the line was found
next # all done for this line
}
{print} # all other lines -> print them
END { # end of file
if(done != 1) # haven't found /option=/ -> add it at the end of output
print "option=value"
}
Run it using
awk -f update.awk < /etc/fdm_monitor.conf > /etc/fdm_monitor.conf.tmp && \
mv /etc/fdm_monitor.conf.tmp /etc/fdm_monitor.conf
or
awk -f update.awk < /etc/fdm_monitor.conf | sponge /etc/fdm_monitor.conf
EDIT:
As a one-liner:
awk '/^option *=/ {print "option=value";d=1;next}{print}END{if(d!=1)print "option=value"}' /etc/fdm_monitor.conf | sponge /etc/fdm_monitor.conf
use awk
awk 'FNR==NR && /configs.*projectname\.conf/{f=1;next}f==0;END{ if(!f) { print "your line"}} ' file file
sed -i 's/^option.*/option=value/g' /etc/fdm_monitor.conf
grep -q "option=value" /etc/fdm_monitor.conf || echo "option=value" >> /etc/fdm_monitor.conf
here is an awk one-liner:
awk -v s="option=value" '/^option/{f=1;$0=s}7;END{if(!f)print s}' file
this doesn't do in-place change on the file, you can however :
awk '...' file > tmpfile && mv tmpfile file
Using sed, you could say:
sed -e '/option=/{s/.*/option=value/;:a;n;:ba;q}' -e 'aoption=value' filename
This would replace the parameter if it exists, else would add it to the bottom of the file.
Use the -i option if you want to edit the file in-place:
sed -i -e '/option=/{s/.*/option=value/;:a;n;:ba;q}' -e 'aoption=value' filename
sed -i '1 h
1 !H
$ {
x
s/^option.*/option=value/g
t
s/$/\
option=value/
}' /etc/fdm_monitor.conf
Load all the file in buffer, at the end, change all occurence and if no change occur, add to the end
The answers using grep are wrong. You need to add an -x option to match the entire line otherwise lines like #text to add will still match when looking to add exactly text to add.
So the correct solution is something like:
grep -qxF 'include "/configs/projectname.conf"' foo.bar || echo 'include "/configs/projectname.conf"' >> foo.bar
Using sed: It will insert at the end of line. You can also pass in variables as usual of course.
grep -qxF "port=9033" $light.conf
if [ $? -ne 0 ]; then
sed -i "$ a port=9033" $light.conf
else
echo "port=9033 already added"
fi
Using oneliner sed
grep -qxF "port=9033" $lightconf || sed -i "$ a port=9033" $lightconf
Using echo may not work under root, but will work like this. But it will not let you automate things if you are looking to do it since it might ask for password.
I had a problem when I was trying to edit from the root for a particular user. Just adding the $username before was a fix for me.
grep -qxF "port=9033" light.conf
if [ $? -ne 0 ]; then
sudo -u $user_name echo "port=9033" >> light.conf
else
echo "already there"
fi
I elaborated on kev's grep/sed solution by setting variables in order to reduce duplication.
Set the variables in the first line (hint: $_option shall match everything on the line up until the value [including any seperator like = or :]).
_file="/etc/ssmtp/ssmtp.conf" _option="mailhub=" _value="my.domain.tld" \
sh -c '\
grep -q "^$_option" "$_file" \
&& sed -i "s/^$_option.*/$_option$_value/" "$_file" \
|| echo "$_option$_value" >> "$_file"\
'
Mind that the sh -c '...' just has the effect of widening the scope of the variables without the need for an export. (See Setting an environment variable before a command in bash not working for second command in a pipe)
You can use this function to find and search config changes:
#!/bin/bash
#Find and Replace config values
find_and_replace_config () {
file=$1
var=$2
new_value=$3
awk -v var="$var" -v new_val="$new_value" 'BEGIN{FS=OFS="="}match($1, "^\\s*" var "\\s*") {$2=" " new_val}1' "$file" > output.tmp && sudo mv output.tmp $file
}
find_and_replace_config /etc/php5/apache2/php.ini max_execution_time 60
If you want to run this command using a python script within a Linux terminal...
import os,sys
LINE = 'include '+ <insert_line_STRING>
FILE = <insert_file_path_STRING>
os.system('grep -qxF $"'+LINE+'" '+FILE+' || echo $"'+LINE+'" >> '+FILE)
The $ and double quotations had me in a jungle, but this worked.
Thanks everyone
Try:
LINE='include "/configs/projectname.conf"'
sed -n "\|$LINE|q;\$a $LINE" lighttpd.conf >> lighttpd.conf
Use the pipe as separator and quit if $LINE has been found. Otherwise, append $LINE at the end.
Since we only read the file in sed command, I suppose we have no clobber issue in general (it depends on your shell settings).
Using only sed I'd suggest the following solution:
sed -i \
-e 's#^include "/configs/projectname.conf"#include "/configs/projectname.conf"#' \
-e t \
-e '$ainclude "/configs/projectname.conf"' lighttpd.conf
s replace the line include "/configs/projectname.conf with itself (using # as delimiter here)
t if the replacement was successful skip the rest of the commands
$a otherwise jump to the last line and append include "/configs/projectname.conf after it
Almost all of the answers work but not in all scenarios or OS as per my experience. Only thing that worked on older systems and new and different flavours of OS is the following.
I needed to append KUBECONFIG path to bashrc file if it doesnt exist. So, what I did is
I assume that it exists and delete it.
with sed I append the string I want.
sed -i '/KUBECONFIG=/d' ~/.bashrc
echo 'export KUBECONFIG=/etc/rancher/rke2/rke2.yaml' >> ~/.bashrc
I needed to edit a file with restricted write permissions so needed sudo. working from ghostdog74's answer and using a temp file:
awk 'FNR==NR && /configs.*projectname\.conf/{f=1;next}f==0;END{ if(!f) { print "your line"}} ' file > /tmp/file
sudo mv /tmp/file file
I am using sed in a script to do a replace and I want to have the replaced file overwrite the file. Normally I think that you would use this:
% sed -i 's/cat/dog/' manipulate
sed: illegal option -- i
However as you can see my sed does not have that command.
I tried this:
% sed 's/cat/dog/' manipulate > manipulate
But this just turns manipulate into an empty file (makes sense).
This works:
% sed 's/cat/dog/' manipulate > tmp; mv tmp manipulate
But I was wondering if there was a standard way to redirect output into the same file that input was taken from.
I commonly use the 3rd way, but with an important change:
$ sed 's/cat/dog/' manipulate > tmp && mv tmp manipulate
I.e. change ; to && so the move only happens if sed is successful; otherwise you'll lose your original file as soon as you make a typo in your sed syntax.
Note! For those reading the title and missing the OP's constraint "my sed doesn't support -i": For most people, sed will support -i, so the best way to do this is:
$ sed -i 's/cat/dog/' manipulate
Yes, -i is also supported in FreeBSD/MacOSX sed, but needs the empty string as an argument to edit a file in-place.
sed -i "" 's/old/new/g' file # FreeBSD sed
If you don't want to move copies around, you could use ed:
ed file.txt <<EOF
%s/cat/dog/
wq
EOF
Kernighan and Pike in The Art of Unix Programming discuss this issue. Their solution is to write a script called overwrite, which allows one to do such things.
The usage is: overwrite file cmd file.
# overwrite: copy standard input to output after EOF
opath=$PATH
PATH=/bin:/usr/bin
case $# in
0|1) echo 'Usage: overwrite file cmd [args]' 1>&2; exit 2
esac
file=$1; shift
new=/tmp/overwr1.$$; old=/tmp/overwr2.$$
trap 'rm -f $new $old; exit 1' 1 2 15 # clean up
if PATH=$opath "$#" >$new
then
cp $file $old # save original
trap '' 1 2 15 # wr are commmitted
cp $new $file
else
echo "overwrite: $1 failed, $file unchanged" 1>&2
exit 1
fi
rm -f $new $old
Once you have the above script in your $PATH, you can do:
overwrite manipulate sed 's/cat/dog/' manipulate
To make your life easier, you can use replace script from the same book:
# replace: replace str1 in files with str2 in place
PATH=/bin:/usr/bin
case $# in
0|2) echo 'Usage: replace str1 str2 files' 1>&2; exit 1
esac
left="$1"; right="$2"; shift; shift
for i
do
overwrite $i sed "s#$left#$right#g" $i
done
Having replace in your $PATH too will allow you to say:
replace cat dog manipulate
You can use sponge from the moreutils.
sed "s/cat/dog/" manipulate | sponge manipulate
Perhaps -i is gnu sed, or just an old version of sed, but anyways. You're on the right track. The first option is probably the most common one, the third option is if you want it to work everywhere (including solaris machines)... :) These are the 'standard' ways of doing it.
To change multiple files (and saving a backup of each as *.bak):
perl -p -i -e "s/oldtext/newtext/g" *
replaces any occurence of oldtext by newtext in all files in the current folder. However you will have to escape all perl special characters within oldtext and newtext using the backslash
This is called a “Perl pie” (mnemonic: easy as a pie)
The -i flag tells it do do in-place replacement, and it should be ok to use single (“'”) as well as double (“””) quotes.
If using ./* instead of just *, you should be able to do it in all sub-directories
See man perlrun for more details, including how to take a backup file of the original.
using sed:
sed -i 's/old/new/g' ./* (used in GNU)
sed -i '' 's/old/new/g' ./* (used in FreeBSD)
-i option is not available in standard sed.
Your alternatives are your third way or perl.
A lot of answers, but none of them is correct. Here is the correct and simplest one:
$ echo "111 222 333" > file.txt
$ sed -i -s s/222/444/ file.txt
$ cat file.txt
111 444 333
$
Workaround using open file handles:
exec 3<manipulate
Prevent open file from being truncated:
rm manipulate
sed 's/cat/dog/' <&3 > manipulate