find a string in every line of a file then replace another string on the same line using bash-CodePudding

here is my script :

#! /bin/bash

# C-band Edit
dqt='"'
str5="polarization=${dqt}2${dqt}"
for x in {3600..4200} do;
    sed -i "/$str5.*$x/s/$x/$((x-600))/" satellites.xml
done

all i want to do is to replace number x with x-600 in between 3600 and 4200 on all lines that have str5 in satellites.xml but the script above gives me syntax error

CodePudding user response：

You don't need to use grep -n to get the line number. You can use a regular expression as the address for a command.

$x-600 won't do subtraction, you need to put it in an arithmetic expression context with $(()).

for x in {3600..4200}
do
    sed -i "/$strAAA.*$x/s/$x/$((x-600))/" file.txt
done

Although running sed 600 times to replace each number is a pretty inefficient way to do it. awk would be a better solution if the number to replace is at a consistent place in the line. E.g. if it's in the 5th field:

awk -v search="$strAAA" '$1 == search { $5 -= 600 }' file.txt > newfile.txt

CodePudding user response：

@Barmar i use bash for sure this is my final state for script

#! /bin/bash

# C-band Edit
dqt='"'
str5="polarization=${dqt}2${dqt}"
for x in {3600..4200} do;
    sed -i "/$str5.*$x/s/$x/$((x-600))/" satellites.xml
done

it gives me syntax error