For example, I have string:
explicit_borders="\"\"''()"
How to calculate based on this string the string with the following result:
distinct_border_chars="\"'()" # contains only unique set of chars
To solve this should I loop over each char in the string and add the char only if the result string does not contain it? Or perhaps is there some simpler bash-specific solution?
Update
I would prefer solutions based on bash built-ins without additional subshell creation (that are using pipes | and command substitution with $()) since they will work better from stand point of performance.
And better without here-strings (<<< ...) since they are creating temporary files which impacts performance.
But actually any solutions can also be considered (especially simple and concise) to have more choice.
Yes it can be done in pure bash, all we need is an associative array to behave as a hash-map and key each character as it occurs in the input string. If the char has appeared once, append it to the final result
#!/usr/bin/env bash
explicit_borders="\"\"''()"
declare -A seen
result=""
for (( i=0; i<${#explicit_borders}; i++ )); do
c="${explicit_borders:$i:1}"
if ! [[ -n "${seen[$c]}" ]]; then
seen[$c]=1
result+="$c"
fi
done
printf -v result '%q' "$result"
printf "%s\n" "${result}"
The %q format specifier at the end causes printf to output the corresponding argument in a format that can be reused as shell input.