I'm cleaning up a government dataset that I need some help on. For the sake of simplicity, let's say the dataset looks like this:

namejob1 is a string variable that has individual names and their job title

namejob1
Tina Lee Firefighter
Border John Doe Patrol
Trevor Secretary Jones
Legislative Aide Samuel Adams-Perez
Jane Smith Highway Police
job2 is a string variable has individual job titles extracted from namejob1

job2
Firefighter
Border Patrol
Secretary
Legislative Aide
Highway Police
I have over 1000 unique values (i.e., job titles) in job2 and the number of unique observations in namejob1 is over a million. namejob1 is messy overall with the job titles sitting in different places (i.e., beginning, middle, end, etc.) and some are even broken up (see, Border John Doe Patrol).

I've been trying to figure out how to subtract job2 from namejob1 into a new variable so I get the desired_name3 (listed below). desired_name3 would be a new string variable that just has individual names (namejob1 - job2)

desired_name3
Tina Lee
John Doe
Trevor Jones
Samuel Adams-Perez
Jane Smith
Any thoughts? I'm trawled the webs and Stack Overflow and I'm not finding anything helpful in subtracting string variables using another string variable as the substring (most are solutions for fixed substrings, like "vs." or parsing using symbols or spaces).

Thanks so much.