I wish to take 10 random lines of file, which is 100 lines long. First, I randomly generate 10 integers between 1 and 100 (inclusive) with

ind=$(shuf -i 1-100 -n 10 | sort -n)

Then, I wish to use sed to extract out those lines (indicated by ind). So I need to re-print the array ind to generate

<ind(1)>p;<ind(2)>p;...;<ind(10)>p

as in

sed -n '<ind(1)>p;<ind(2)>p;...;<ind(10)>p' ~/orig.txt > ~/short.txt

How may I do this?

In your example, ind is a string, not an array. You must use:

ind=($(shuf -i 1-100 -n 10 | sort -n))

to make ind to be array (in shell support array like bash, zsh, ksh).

Simply, you can try:

$ printf '%spn' $(shuf -i 1-100 -n 10 | sort -n) > /tmp/short.sed
$ sed -n -f /tmp/short.sed < orig.txt > short.txt

An awk based solution (without need of sorting):

awk 'NR==FNRa[$1];next NR in a' <(shuf -i 1-100 -n 10) ~/orig.txt > ~/short.txt

And a pure GNU awk variant (without need of other external processes):

awk '
 BEGIN srand(); do a[int(100*rand()+1)]; while (length(a)<10) 
 NR in a
' ~/orig.txt > ~/short.txt

Instead of trying to go the long way, creating a sed script to extract the lines from your file, just use shuf directly on the file:

shuf -n 10 ~/short.txt

If you want the lines in the order that they are found in the original file:

cat -n ~/short.txt | shuf -n 10 | sort -n | cut -f 2-

This enumerates the lines in the file, extracts 10 lines from it (in random order), sorts the extracted lines, and removes the numbering.

U could...

shuf -i 1-100 -n 10 | sort -n | sed 's/$/p;/' | sed -nf - orig.txt >short.txt

or, under bash

sed -nf <(sed 's/$/p;/' <(sort -n < <(shuf -i 1-100 -n 10))) <orig.txt >short.txt