How does SED work with variables?

Question

linux

I execute the command

sed -i "s/\(\([^,]\+,\)\{9\}\)/\1$((RANDOM%1000000+8243678))/" file.txt

According to it, he puts a random number after the 9th comma in each line of the file.
But for some reason, apparently to understand which I do not have enough knowledge, Sed, instead of getting the value of the variable RANDOM% 1000000 + 8243678 for each line, takes and puts down the same (apparently the first received) value of this variable. Perhaps this is a feature of how Sed works with the -i option. Who can tell whether it is possible through Sed in this case to implement and how to set the value of the RANDOM variable for each unique line.

Answer 1

The RANDOM function is executed once when it is called, sed calls this function once when opening a file, so you need to loop it, like this:

cat file.txt | while read line; do echo $line | sed "s/\(\([^,]\+,\)\{9\}\)/\1$((RANDOM%1000000+8243678))/"; done

Answer 2

SED does not work with this variable in any way. When processing a command, the shell replaces the variable with a value that has already been calculated before even running sed itself.
As a result, sed does something like:
Konkase is right in advising you to use a loop.

Answer 3

you have one thread running and it contains one variable $RANDOM
try to make a loop for each line with the assignment of this variable. in fact, for each cycle it will be generated anew

Answer 4

the variable is taken in curly braces
example
sed -i "/${controlnumber}/d" ./000