Complicated string replacement in files - string

Question

So, I have some strings to replace in some data files for a game I'm working on. I'm fine with using either sed or awk. I need help programming the bash script to replace this, so here's a list of things I want to replace. (Keep in mind that I will be using X in brackets in place of numbers, also, for strings 1 and 2 there is a space after the string I need to replace, which I need to replace the space as well.)
Replace every instance of \\P[x]: with \\n<\\P[x]>\\pf[x]
Replace every instance of \\N[x]: with \\n<\\N[x]>\\af[x]
Replace every instance of "volume":80 with "volume":90
Replace every instance of ... with \\..\\..\\..\\.
Replace every instance of "]},{"code":401,"indent":0,"parameters":[" with "]},{"code":401,"indent":0,"parameters":[" (There is a space at the beginning of the string I need to replace, I want to move that to the end of the string.)
Here's what I have tried:
sed -e 's/"volume":80/"volume":90/g' -e 's///g' -e 's/\ \"]},{\"code\":401,\"indent\":0,\"parameters\":[\"/\"]},{\"code\":401,\"indent\":0,\"parameters\":[\"\ /g' < ${file} > ${outdir}/${filename}.${extension}
I get this error: `sed: -e expression #3, char 109: unterminated s' command
And here's a script I created, can't figure out how to get it to work right:
#!/bin/bash
direct=./1
outdir=./2
mkdir -p $direct
mkdir -p $outdir
for file in ${direct}/*.*; do
filepath=$(basename "$file")
extension="${filepath##*.}"
filename="${filepath%.*}"
strb='"volume":80'
strr='"volume":90'
echo 1 && sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
strb='...'
strr='\\..\\..\\..\\.'
echo 2 && sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
num=0
while [ "$num" -le "99" ]; do
strb='\\P['"$num"']\: '
strr='\\n<\\P['"$num"']>\\pf['"$num"']'
sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
strb='\\N['"$num"']\: '
strr='\\n<\\N['"$num"']>\\af['"$num"']'
sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
num=$(echo $num+1 | bc -l)
done
strb=' "]},{"code":401,"indent":0,"parameters":["'
strr='"]},{"code":401,"indent":0,"parameters":[" '
echo 5 && sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
strb='Move1'
strr='Move'
echo 6 && sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
strb='Move'
strr='Move1'
echo 7 && sed -e "s#\"${strb}\"#\"${strr}\"#g" -i ${file}
echo ${filename}.${extension}
done
On the fifth instance of sed, I get unterminated s' command

Answer 1

I think you get unterminated s' command on the fifth instance of sed in your script because your s' command is terminated with double-quotes, but your $strb and $strr strings also have unescaped double quotes in them. This is a problem for your entire script, but luckily, the fifth instance happens to have an odd number of double quotes, causing a syntax error and revealing the problem. Since none of your strings have single-quotes, using only single-quotes to terminate sed and wrapping all of your variables in double-quotes to preserve their literals is a safe bet, e.g.,
sed -e 's#'"${strb}"'#'"${strr}"'#g' -i ${file}
You also have some [] and some . in your strings. These have a special meaning to sed, so you should escape those (for the pattern, not the replacement), e.g.,
change
strb='...'
to
strb='\.\.\.'
and change
strb=' "]},{"code":401,"indent":0,"parameters":["'
to
strb=' "\]},{"code":401,"indent":0,"parameters":\["'
See https://unix.stackexchange.com/a/33005/65267 for more on characters with special meaning to sed.