escape string and make it unix ready

Question

I would like to take a string that may have multiple spaces in it and do the following:

1) Replace whitespace with an underscore

2) Remove any characters that are not A-Z or 0-9

3) Make the result all lowercase

Then be able to use the result as a variable. Any ideas?

Answer 1

I think tr can do what you want.

variable=$(echo "${input}" | tr A-Z a-z | tr [:blank:] _ | tr -cd [:alnum:]_)

Explanation:

tr A-Z a-z - Translate upper cased letters to lower cased.

tr [:blank:] _ - Translates blank spaces to underscores.

tr -cd [:alnum:]_ - Delete all characters that are not alphanumeric, or underscores.

NOTE: If you want to remove underscores prior to converting spaces to underscores, tr -cd _ could be added near the beginning of the pipe chain.

Answer 2

Here's another approach using sed:

oldvar="HELLO MY BABY @$#@$ HI"

newvar=`echo $oldvar | sed -e "s/[A-Z]/\l&/g" -e "s/[^a-z0-9]/_/g"`

results in:

hello_my_baby__0___hi

Answer 3

newvar=`echo "$oldvar" | awk '{gsub(/[ \t]+/,"_") ; gsub(/[^A-Z0-9]+/,"") ; print tolower($0)}' `