Removing list of comma separated words from a sentence

Question

I have two variables as follows :

sentence="name string,age int,address string,dob timestamp,job string"
ignore="age int,dob timestamp"

Basically I need to iterate through the comma separated variable $ignore and remove each word from the above varibale $sentence.

After performing this the output sentence should be as below:

echo $outputsentence
name string,address string,job string

Should I create an array for words to be ignored and iterate through it an perform a sed operation? Is there any other way around?

Answer 1

This is a situation that requires sets: you want to know which members of set A are not present in set B.

For this we have a beautiful article Set Operations in the Unix Shell that describes all of them.

If you want to check the intersection of sets, say:

$ comm -12 <(tr ',' '\n' <<< "$sentence" | sort) <(tr ',' '\n' <<< "$ignore" | sort)
age int
dob timestamp

For complement, use comm -23:

$ comm -23 <(tr ',' '\n' <<< "$sentence" | sort) <(tr ',' '\n' <<< "$ignore" | sort)
address string
job string
name string

Note tr ',' '\n' <<< "$var" | sort just splits the ,-separated strings in slices. Then, <( ) is a process substitution.

Answer 2

With GNU sed:

pattern=$(sed "s/,/|/g" <<< "$ignore")
outputsentence=$(sed -r 's/('"$pattern"'),*//g' <<< "$sentence")

The first sed command replace all , with an alternation operator | in the ignore list.

This result is used as a pattern in to remove the strings from $sentence.