CODEWITHSUNDEEP
stringbashcomparecontains
Алдар
Алдар

Reputation: 175

Compare array elements and determine which element's string is the part of another in bash

I have an array: 

src_dir=( /etc /var/www /var /var/html /home/user)

I want output like this:

/etc
/var contains /var/www /var/html
/home/user

I've tried so far:

src_dir=( /etc /var/www /var /var/html /home/user)
declare exclude
for src in ${src_dir[@]}; do
    if ! [[ " $src " =~ $exclude ]]; then
        exclude+=("$src")
    fi
done

Upvotes: 1

Views: 47

Answers (1)

pjh
pjh

Reputation: 8064

Try this Shellcheck-clean code:

#! /bin/bash -p

src_dir=( /etc /var/www /var /var/html /home/user)

for src in "${src_dir[@]}"; do
    # Determine if $src contains other elements ($is_container), and which
    # elements contain it ($containers)
    is_container=0
    containers=()   # Elements that this element is part of
    for src2 in "${src_dir[@]}"; do
        if [[ $src2 == "$src" ]] ; then
            # Don't count an element as being part of itself
            continue
        elif [[ $src == *"$src2"* ]]; then
            is_container=1
            break
        elif [[ $src2 == *"$src"* ]] ; then
            containers+=( "$src2" )
        fi
    done

    # Skip elements that contain other elements.
    # Print elements elements that don't contain other elements, and any
    # elements that contain them.
    if (( is_container )) ; then
        continue
    elif (( ${#containers[*]} > 0 )) ; then
        printf '%s is contained in' "$src"
        printf ' %q' "${containers[@]}"
        printf '\n'
    else
        printf '%s\n' "$src"
    fi
done
  • Regular expression matching (=~) has some oddities in Bash, and is best avoided. Simple pattern matching with [[ ... == ... ]] is sufficient here.
  • Elements aren't listed as being contained in other elements if they contain any elements. So if you've got (/home /home/user /home/user/dir) then /home/user will not be listed as being contained in /home/user/dir. That may not be what you want.
  • Elements are compared as pure strings. No attempt is made to check for filesystem path containment. So, for instance, /home/user is considered to be contained in /home/user2. That may not be what you want.
  • If an element contains two other elements, it will be listed as containing both of them. So both /home and /user will be listed as being contained in /home/user. That may not be what you want.

Upvotes: 1

Related Questions