Why wont sed remove line from file?

Question

I\'m trying to remove a specific line from a file and then append the edited line to the file. I get last part right but my sed command to remove the old line is not working.

Answer 1

$ is a special character in sed meaning "end of line". If you want to match the literal $userinput at the begging of a line use see '/^\$userinput/d'

If you want your shell to substitute the variable $userinput with its content, use double quotes instead of single quotes:

sed "/^$userinput/d" file1.txt > file2.txt

or better yet

sed "/^${userinput}/d" file1.txt > file2.txt

That will remove lines starting with the value of ${userinput}, if you want to only delete lines that are exactly the value of ${userinput}, use

sed "/^${userinput}$/d" file1.txt > file2.txt

works for me:

╰─ ☭ cat file 
bar
foobar
foo
asdf
qwer
foo
asdf
╰─ ☭ cat script.sh 
#!/bin/bash
sed "/^${1}$/d" file > otherfile
╰─ ☭ ./script.sh foo
╰─ ☭ cat otherfile 
bar
foobar
asdf
qwer
asdf

Answer 2

i did try that and found that the variable value wasnt properly replaced by the shell in the sed statement . so the sed command instead of searching for the value as the pattern was doing a search with the variable name as pattern

ie it was trying to search $userinput as pattern @ EOL

this was due to the presence of single quotes , try your command again but replace them with double quotes : it worked im my sys.

code ::

Nitin@Kaizen ~>  cat new

#!/bin/bash

echo "the input file :"
cat INPUT2.txt ;

echo "enter line to remove" ;
read rem ;

date >> sed.log ;

sed "/$rem/d" INPUT2.txt >> sed.log;

echo "the output file :" ;
cat sed.log;

run ::

Nitin@Kaizen ~ >  ./new

the input file :
aa1
chap1
mk.t
temp.txt
z1
z2

enter line to remove
temp.txt

the output file :
Thu Mar 28 08:18:07 IST 2013
aa1
chap1
mk.t
z1
z2

Answer 3

figured out how to fix it.

update(){
read -p "Enter the course number of the course you would like to update: " updateInput
grep -v ^$updateInput my_course.txt>updatedfile.txt
grep ^$updateInput my_course.txt>editedline.txt
if [ $? -eq 0 ]
  then
    while(true)
      do
            read -p "Status: " status
            if [ $status -eq 0 ]
            then
              break
            elif [ $status -eq 1 ]
            then
              break
            else
              printf "Enter either a 0 or 1\n"  
            fi  
      done  
    read -p "Grade: " grade  
    while(true)  
      do  
        case $grade in  
              A)  
                break  
              B)
              break
                ;;
              C)
                break
                ;;
              *)
                read -p "Enter either an A, B, C: " grade
            ;;
            esac
      done
      awk -F, '
        /^CSC/{printf$1","$2","$3"," >> "updatedfile.txt"}' editedline.txt
      printf $status","$grade"\n" >> updatedfile.txt
      cat updatedfile.txt > my_course.txt
      rm updatedfile.txt
      rm editedline.txt
  else
    printf "This course doesn't exist in your courses list.\n"
fi
}  

In the beginning of the code I use

grep -v regexexperssion file1>file2 

to get all lines in the file without the one to be removed. Then I use grep to get the actual line to be edited from the file.

grep regexexperssion file1>file3  

At the bottom in the awk I perform the edits using the user generated variables on the line in file3 and then I append that to file1, which was generated with the

grep -v regexexpression file1>file2   

command

NOTE the above explanation does not match the actual code used above, BECAUSE i thought it would better for people viewing this letter for a more general explanation. Also I know that there is probably a much better way to solve this problem and that the code above could probably be shortened. Any input on how to improve this would be greatly appreciated. Thanks to all those who helped me.

Answer 4

In this case you need do this because $ means end of line regular expressions.

Try:

sed "/^${userinput}/d" file1.txt >> file2.txt

Also >> will append to the file.