Sed replacing last comma fails - sed

Question

I'm working on a sed script that takes a bunch of lines and turns them into an argument list for matlab (single quoted, comma separated).
It's working well so far:
[script to generate list] | sed -n "s#\(.*$\)#'\1',#p#" | tr '\n' ' '
But this leaves me with a trailing comma.
By testing, I can remove it with
[list of comma separated values] | sed -n 's#,$##p#'
but, when putting it all together:
[script to generate list] | sed -n "s#\(.*$\)#'\1',#;s#,$##p#" | tr '\n' ' '
Outputs nothing.
I feel like it has something to do with not having a p in the first line of the sed script, but I don't want it to print those values, I want them sent to the next line in the script (isn't that the default?)
Edit:
[script to generate list] Outputs a list of directories, for example:
./work/matlab_stun_gun/tex/fullTest.pdf
./Downloads/Howfar(tetra2) fixed.pdf
./work/savdocs/win_tests/tex/texReport.pdf
./Downloads/AcademicAudit.pdf
./work/matlab_stun_gun/report.pdf
./Downloads/PMB_4DVMC.pdf
./work/savdocs/win_tests/tex/mouseHeatMap.pdf
./Downloads/Geometry.pdf
./work/savdocs/win_tests/tex/mouseHeatMap.pdf
./work/matlab_stun_gun/tex/fullTest.pdf
The list generator is just find . -name "*.pdf" | pickl -n 10, adjusted for file type/ number etc. This is going to become a general purpose script.
Expected output would be :
'./work/savdocs/win_tests/tex/mClickss.pdf', './Downloads/Howfar(tetra2) fixed.pdf', './Downloads/MedPhys_defDOSXYZ.pdf', './Downloads/MedPhys_defDOSXYZ.pdf', './report.pdf', './work/savdocs/win_tests/tex/cSwitchs.pdf', './tex/zoomIn.pdf', './tex/fullTest.pdf', './temp/tex/zoomIn.pdf', './tex/zoomIn.pdf'
(Note the lack of trailing comma)

Answer 1

You are experiencing a multi-faceted problem here, in the sense that each of your attempts has something wrong with it.
Starting with [list of comma separated values] | sed -n 's#,$##p#', keep in mind that tr effectively makes your separator ', ' (comma-space) instead of just ',' comma. This means that you will output nothing from the second sed expression. You can fix that by matching with sed -n 's#, $##p#'. If you insist on using the -n flag, that is the correct solution. In full:
[script to generate list] | \
sed -n "s#\(.*$\)#'\1',#p#" | \
tr '\n' ' ' | \
sed -n 's#, $##p#'
The problem with your combination attempt, [script to generate list] | sed -n "s#\(.*$\)#'\1',#;s#,$##p#" | tr '\n' ' ', is that you need to apply tr before you remove the trailing commas. Even if this were to print anything, you would be adding a comma, stripping it off immediately on each line, and then replacing newlines with spaces. The correct order is already shown above.
Multiple commands in sed should be specified using the -e flag. They pipe the result of one command into another, equivalently to using pipes, but much more efficiently. To get sed -n "s#\(.*$\)#'\1',#;s#,$##p#" to print, rephrase it like:
sed -n -e "s#\(.*$\)#'\1',#" -e "s#, $##p#"
This is of course going to strip off the commas as soon as you add them to each line, but it shows the correct syntax for doing so.
Further Improvements
You probably don't need to use the -n flag for sed (and consequently) the /p/ flag for the s command. The -n flag is only useful if you only want to print matches, but you want to print everything, so it does not apply to you.
You also don't need an explicit capture group since you can use the \0 replacement to get the entire match, not just the group. Here is an example:
[script to generate list] | sed "s/.*/'\0'" | tr '\n' ' ' | sed 's/, $//'
Finally, there are alternatives to removing the trailing bits of the string without starting a subprocess, especially since you are already enclosing your expression in $(...):
RESULT=$([script to generate list] | sed "s/.*/'\0'" | tr '\n' ' ')
RESULT=${RESULT%, }
OR
RESULT=${RESULT::-2}