I have two commands. The first, when stored in a script variable, gives output like this:

one two three four five

The second also gives a list, but some of the items may be missing that were in the first command:

one three five

I want my script to do something if an item is in the first command but not the second. None of the items will have spaces (they tend to be kabab-format). How can I do this in Bash?

CodePudding user response：

Regarding my comment ^[1], I'd tackle this like following:

• Loop over res1
  • If current value does not exist in res2
    • Preform action
    • Stop for if needed

#!/bin/bash

res1=(one two three four five)
res2=(one three five)

for value in "${res1[@]}"; do
    if [[ ! "${res2[*]}" =~ "${value}" ]]; then

        # Do action
        echo "'$value' does not exist in res2"

        # Possibly stop for loop
        break
    fi
done

With the break, this will show:

'two' does not exist in res2

Without the break, it will show:

'two' does not exist in res2
'four' does not exist in res2

CodePudding user response：

One approach using the current variables, and relying on the fact that individual values do not contain embedded white space:

$ var1='one two three four five'
$ var2='one three five'
$ comm -23 <(printf "%s\n" ${var1} | sort) <(printf "%s\n" ${var2} | sort)
four
two

NOTE: do not wrap the ${var1} and ${var2} references in double quotes, ie, we want word splitting to occur when feeding the printf calls

Another idea using an associative array to track unique values:

var1='one two three four five'
var2='one three five'

unset      arr
declare -A arr

for f in ${var1}          # use ${var1} values as indices for arr[]
do
    arr[${f}]=1           # '1' has no meaning other than to fill requirement of assigning a value in order to create the array entry
done

for f in ${var2}          # delete ${var2} indices from arr[]
do
    unset arr[${f}]
done


for i in "${!arr[@]}"     # display arr[] indices that remain
do
    echo "${i}"
done

# one-liners (sans comments)

for f in ${var1}; do arr[${f}]=1; done
for f in ${var2}; do unset arr[${f}]; done
for i in "${!arr[@]}"; do echo "${i}"; done

This generates:

two
four

NOTES:

• again, do not wrap the ${var1} and ${var2} references in double quotes, ie, we want word splitting to occur
• if so inclined OP could perform the same add/remove array operations in a single awk script
• the first loop (populating arr[] from ${var1}) will eliminate duplicates from ${var1}, eg, var1='one one one' would lead to a single array entry: arr[one]=1