CODEWITHSUNDEEP
bashshellsed
ilium007
ilium007

Reputation: 589

Delete specific line with sed without deleting other lines with same prefix

I am trying to delete lines of a file using sed but I don't want to delete these lines unless the search string matches exactly. ie.

#!/bin/bash
set -x
file="list1"
index=0

while read line ;
  do
    if [ ! "$line" == "@@@" ];
      then
        MYARRAY[$index]="$line"
        #index=$(($index+1))
        let index++
    else
      break
    fi
done < $file

#echo "MYARRAY is: ${MYARRAY[*]}"
#echo "The file: ${index}"

index1=0
for i in "${MYARRAY[@]}"
do
  lineNumber=$((index1 + 1))
  sed -i "/${MYARRAY[$index1]}/d" "$file"
  let index1++
done

sed -i "/@@@/d" "$file"
#remove empty lines
sed -i '/^$/d' "$file"

So I have a test file (list1):

line1
line2
line3
line4
line5
line6
@@@
line1_1
line2_1

And after the script has run I end up with all lines deleted because "line1" and line2" are also matching "line1_1" and "line2_1"

Any help would be appreciated!!

Upvotes: 1

Views: 1546

Answers (1)

Jamey Sharp
Jamey Sharp

Reputation: 8511

If you anchor your sed pattern match to the beginning and end of the line, then it won't also match substrings. Like so:

sed -i "/^${MYARRAY[$index1]}\$/d" "$file"

(Escaping the '$' used to anchor to the end of the line since it's in a double-quoted string.)

edit: That said, it looks to me like you can replace your entire script with this one-line sed program:

sed '1,/@@@/d; /^$/d'

That deletes from the beginning of the file to the first line containing '@@@', then also deletes blank lines, and does it all in a single pass.

Upvotes: 1

Related Questions