Compare two Arrays in KSH and output the difference

Question

I am not extremely familiar with KSH (Actually just started using it) and I am having problems with trying to create a script that will essentially compare two arrays that have been stored and then output the difference.

Running AIX 6.1 with KSH Version M-11/16/88f

Here is what I have thus far:

#!/bin/ksh
set -A dfArray $(df | awk '{print $7}' | grep -v Mounted)
set -A dsmArray $(cat dsm.sys | grep DOMAIN | awk '{ s=""; for (i = 2; i <= NF; i++) s = s $i " "; print s }')

MAX=$((${#dsmArray[*]}-1))

for a in ${dfArray[*]}
       do
       COUNT=0
       set -A UNMATCHED
       for b in ${dsmArray[*]}
               do
               (( $a != $b )) && \
               UNMATCHED[$COUNT]=$a && \
               let COUNT+=1
       done
       (( ${#UNMATCHED[*]} != $MAX )) && \
       print ${UNMATCHED[0]}
       unset UNMATCHED
done

Here are what the two arrays might look like: dfArray:

/
/usr
/var
/tmp
...

dsmArray:

/home
/opt
/usr
...

When I run the current script I get the following error:

./checkBackup.sh[22]:  / != /home : syntax error

What am I doing wrong? I am sure it's probably something simple.

Thanks!

Solution

You get an error because you're attempting to do arithmetic equality with string values.

Here are 2 ways to check whether the elements of dfArray are in dsmArray

set -A dfArray / /usr /var /tmp ...
set -A dsmArray /home /opt /usr ...

for a in "${dfArray[@]}"; do
  in=false
  for b in "${dsmArray[@]}"; do
    if [[ $a == $b ]]; then
      echo "$a is in dsmArray"
      in=true
      break
    fi
  done
  $in || echo "$a is not in dsmArray"
done

/ is not in dsmArray
/usr is in dsmArray
/var is not in dsmArray
/tmp is not in dsmArray
... is in dsmArray

Or, flattening the dsmArray and avoiding the inner loop:

for a in "${dfArray[@]}"; do
  # all quotes and spaces required below
  if [[ " ${dsmArray[*]} " == *" $a "* ]]; then
    echo "$a is in dsmArray"
  else
    echo "$a is not in dsmArray"
  fi
done

/ is not in dsmArray
/usr is in dsmArray
/var is not in dsmArray
/tmp is not in dsmArray
... is in dsmArray

This solution keys on this part: [[ " ${dsmArray[*]} " == *" $a "* ]]

1. " ${dsmArray[*]} "
   • when quoted, "${ary[*]}" results in a single string composed of the individual elements of the array joined on the first character of $IFS
   • the default of $IFS is 3 characters: space, tab, newline
   • with the leading and trailing quotes, we get the string " /home /opt /usr ... "
2. *" $a "*
   • inside double brackets, the == operator is actually a pattern matching operator, not strictly equality
   • our pattern is:
     • zero or more characters, followed by
     • a space, followed by
     • the value of $a, followed by
     • a space, followed by
     • zero or more characters

So, for each value of dfArray, we are checking to see if it appears as a space-separated word within the flattened dsmArray string.

This breaks down with the array elements can contain spaces. Then you need to assign a new value to IFS and things get messy(-er).

---

To get an array of common elements, I'd do

typeset -i n=0
set -A common
for a in "${dfArray[@]}"; do
  if [[ " ${dsmArray[*]} " == *" $a "* ]]; then
    let n+=1
    common[n]=$a
  fi
done

echo "common"
printf "%s\n" "${common[@]}"
echo "common with index"
typeset -i i=1
while (( i <= n )); do echo "$i  ${common[i]}"; ((i+=1)); done

common
/usr
...
common with index
1  /usr
2  ...

Comments

[Christopher Stanley]

Ahh, duh - That was a silly mistake. Your solution worked perfectly for me. Thanks for your help! Could you explain what is going on when you flatten the dsmArray? I understand your first solution with the nested for statements. Trying to understand your second solution.

[Angel Todorov]

Oh, there's actually an error there, need " ${dsmArray[*]} " -- explanation will be in answer

[Christopher Stanley]

It would appear that ${dfArray[@]} worked too. Is this expected?

[Christopher Stanley]

In fact, changing it to ${dfArray[*]} causes it to output everything, not just the differences.

[Angel Todorov]

I wouldn't have thought so, but I guess the double brackets do some magical things. You can see how the two array expansions differ: print ">%s<\n" " ${dsmArray[*]} " versus print ">%s<\n" " ${dsmArray[@]} "

[Christopher Stanley]

Alright, well the print command on both of those arrays with * and @ output the same thing. Since using [@] on both works for me I will keep it like that. Just was curious as to why, guess I wont question it :P Thanks again for your help!