In an bash shell script I am receiving a list of argument in $@ which are : a b c d a c e Depending of the argument I need to do something specific with a case structure. But I only want to do it for a b c d e
I should only use bash and not any other language...I can use awk for example
for argument in "$@"
do
case $argument in
a)
.....
Have tried many things but without success
Any help much appreciated
Use an associative array to track arguments you've already seen. Note that this requires bash 4.0 or later; the 3.2.x release that Apple ships is too old, as it only supports numerically-indexed arrays (declare -a, but not declare -A).
#!/usr/bin/env bash
case $BASH_VERSION in ''|[0-3].*) echo "ERROR: Bash 4.0+ required" >&2; exit 1;; esac
declare -A seen=( )
declare -a deduped=( )
for arg in "$@"; do # iterate over our argument list
[[ ${seen[$arg]} ]] && continue # already seen this? skip it
seen[$arg]=1 # mark as seen going forward...
deduped+=( "$arg" ) # ...and add to our new/future argv
done
set -- "${deduped[@]}" # replace "$@" with contents of deduped array
deduped, you can set argv from the keys of seen with set -- "${!seen[@]}".You could print, then sort unique, then print quoted the list and re-eval into arguments. With GNU tools:
set -- a b c d a c e
tmp=$(printf "%s\0" "$@" | sort -uz | xargs -0 printf "%q ")
eval set -- "$tmp"
xargs calls /usr/bin/printf, it isn't guaranteed to support %q; that's specific to the shell-builtin printf in bash and ksh (unless it's a recent coreutils addition and I haven't been keeping up, but even then that would be GNU-centric).arr=( ); while IFS= read -r -d '' item; do arr+=( "$item" ); done < <(printf '%s\0' "$@" | sort -uz); set -- "${arr[@]}", that should work.eval "set -- $tmp" over eval set -- "$tmp" just for the sake of being more clear to the code's readers about how eval works, even though they're functionally identical in present circumstances).