I have a variable pattern. And I want to match pattern in file and if pattern is matched then line should be deleted.
I tried with:
sed '/$pattern/d' file.txt
But it doesn't work.
Please give me guidence for the same.
Thanks.
Just do that:
sed /$pattern/d file.txt
The quotes were transforming your variable in a string. Then you need to remove that.
And if you need to to write the changes in the file, just add -i
sed -i /$pattern/d file.txt
Related
I have a file:
dynamicclaspath.cfg
VENDOR_JAR=/clear-as-1-d/apps/sterling/jar/struts/2_5_18/1_0_0/log4j-core-2.10.0.jar
VENDOR_JAR=/clear-as-1-d/apps/sterling/jar/log4j/2_17_1/log4j-core-2.10.0.jar
I want to replace any occurrence of log4j-core* with log4j-core-2.17.1.jar
I tried this but I know I'm missing a regex:
sed -i '/^log4j-core/ s/[-]* /log4j-core-2.17.1.jar/'
With your shown samples please try following sed program. Using -E option with sed to enable ERE(extended regular expressions) with it. In main program using substitute option to perform substitution. Using sed's capability to use regex and store matched values into temp buffer(capturing groups). Matching till last occurrence of / and then matching log4j-core till jar at last of value. While substituting it with 1st capturing group value(till last occurrence of /) followed by new value of log4j as per OP's requirement.
sed -E 's/(^.*\/)log4j-core-.*\.jar$/\1log4j-core-2.17.1.jar/' Input_file
Using sed
$ sed -E 's/(log4j-core-)[0-9.]+/\12.17.1./' input_file
VENDOR_JAR=/clear-as-1-d/apps/sterling/jar/struts/2_5_18/1_0_0/log4j-core-2.17.1.jar
VENDOR_JAR=/clear-as-1-d/apps/sterling/jar/log4j/2_17_1/log4j-core-2.17.1.jar
It depends on possible other contents in your input file how specific the search pattern must be.
sed 's/log4j-core-.*\.jar/log4j-core-2.17.1.jar/' inputfile
or
sed 's/log4j-core-[0-9.]*\.jar/log4j-core-2.17.1.jar/' inputfile
or (if log4j-core*.jar is always the last part of the line)
sed 's/log4j-core.*/log4j-core-2.17.1.jar/' inputfile
sed -i s'#2.10.0.jar$#2.17.1.jar#'g file
That seems to work.
I have a configuration file (gpsd.default) containing data with the following format:
# If you must specify a non-NMEA driver, uncomment and modify the next line
GPSD_SOCKET="/var/run/gpsd.sock"
GPSD_OPTIONS=""
GPS_DEVICES=""
I am making a change on the file with sed:
sed -i 's/^GPS_DEVICES="".*/GPS_DEVICES="dev/ttyUSB1"/' /etc/default/gpsd.default
or
sed -i '4s/^.*/GPS_DEVICES="dev/ttyUSB1"/' /etc/default/gpsd.default
The above sed command returns error:
sed: bad option in substitution expression
Because the new line contains "/" in its expression.
How to update my sed command to make it work?
This is because you are using a regex containing /, which is the same character sed uses as delimiter.
Just change the sed delimiter to another one, for example ~:
sed -i 's~^GPS_DEVICES="".*~GPS_DEVICES="dev/ttyUSB1"~' /etc/default/gpsd.default
By the way, since you are changing files in /etc, you may want to use -i.bak, so that the original file gets backed up. It is a good practice to prevent loss of important information.
You should update your sed command to this.
sed -i 's/^GPS_DEVICES=\"\".*/GPS_DEVICES=\"dev\/ttyUSB1\"/' /etc/default/gpsd.default
My question is probably rather simple. I'm trying to replace sequences of strings that are at the beginning of lines in a file. For example, I would like to replace any instance of the pattern "GN" with "N" or "WR" with "R", but only if they are the first 2 characters of that line. For example, if I had a file with the following content:
WRONG
RIGHT
GNOME
I would like to transform this file to give
RONG
RIGHT
NOME
I know i can use the following to replace any instance of the above example;
sed -i 's/GN/N/g' file.txt
sed -i 's/WR/R/g' file.txt
The issue is that I want this to happen only if the above patterns are the first 2 characters in any given line. Possibly an IF statement, although i'm not sure what the condition would look like. Any pointers in the right direction would be much appreciated, thanks.
just add the circumflex, remove g suffix (unnecessary, since you want at most one replacement), you can also combine them in one script.
sed -i 's/^GN/N/;s/^WR/R/' file.txt
Use the start-of-string regexp anchor ^:
sed -i 's/^GN/N/' file.txt
sed -i 's/^WR/R/' file.txt
Since sed is line-oriented, start-of-string == start-of-line.
I have a configuration file (gpsd.default) containing data with the following format:
# If you must specify a non-NMEA driver, uncomment and modify the next line
GPSD_SOCKET="/var/run/gpsd.sock"
GPSD_OPTIONS=""
GPS_DEVICES=""
I am making a change on the file with sed:
sed -i 's/^GPS_DEVICES="".*/GPS_DEVICES="dev/ttyUSB1"/' /etc/default/gpsd.default
or
sed -i '4s/^.*/GPS_DEVICES="dev/ttyUSB1"/' /etc/default/gpsd.default
The above sed command returns error:
sed: bad option in substitution expression
Because the new line contains "/" in its expression.
How to update my sed command to make it work?
This is because you are using a regex containing /, which is the same character sed uses as delimiter.
Just change the sed delimiter to another one, for example ~:
sed -i 's~^GPS_DEVICES="".*~GPS_DEVICES="dev/ttyUSB1"~' /etc/default/gpsd.default
By the way, since you are changing files in /etc, you may want to use -i.bak, so that the original file gets backed up. It is a good practice to prevent loss of important information.
You should update your sed command to this.
sed -i 's/^GPS_DEVICES=\"\".*/GPS_DEVICES=\"dev\/ttyUSB1\"/' /etc/default/gpsd.default
How can I replace a string but only in the first line of the file using the program "sed"?
The commands s/test/blah/1 and 1s/test/blah/ don't seem to work. Is there another way?
This might work for you (GNU sed):
sed -i '1!b;s/test/blah/' file
will only substitute the first test for blah on the first line only.
Or if you just want to change the first line:
sed -i '1c\replacement' file
This will do it:
sed -i '1s/^.*$/Newline/' textfile.txt
Failing that just make sure the match is unique to line one only:
sed -i 's/this is line one and its unique/Changed line one to this string/' filename.txt
The -i option writes the change to the file instead of just displaying the output to stdout.
EDIT:
To replace the whole line by matching the common string would be:
sed -i 's/^.*COMMONSTRING$/Newline/'
Where ^ matches the start of the line, $ matches the end of the line and .* matches everything upto COMMONSTRING
this replaces all matches, not just the first match, only in the first line of course:
sed -i '1s/test/blah/g' file
the /g did the trick to replace more than one matches, if any exists.