i extract a number from a text and tried using
bc -l << $num1/$num2 which kept giving error.

i found one of the number has contained invisible garbage char.
i only found about by
echo $num1 > file.txt
and
hexdump -C file.txt

which prints out ascii code of each byte in the text file.
i can see it contains char like 1b (escape), 6d (letter m), 5d (bracket) where visible char only contains digits and dot for numeric calculation.

I have tried myriad of methods with cut, tr with non-printable chars like [:space:], [:cntrl:] but does not work.

I think I need to figure out how to use sed in this case but by specifying the ascii code.

probably:

invisible char. Ascii usually in range of 1F and less than it and 7f and more, it depends on your precision, so removal of them is

1 members found this post helpful.

keeps a - z A - Z 0 - 9 gets rid of the rest.

sed with [[:cntrl:]] would be my preferred - simple and elegant.

Use tr 'backwards':

removes every non-alphanumeric character.