Question 4Data and SQL
Two HR systems list the same people under different IDs. How do you match them?
Matching records across systems is routine FDE work and easy to get quietly wrong.
“Data Engineering & Integration: Proficiency in complex SQL, ETL/ELT pipeline design, and experience building system integrations connecting AI solutions to enterprise and telecom data sources.”Senior Engineer, Forward Deployment · T-Mobile
The answerFour steps
- 1
Find the fields both systems trust: work email, employee number, name plus start date
Say it like this
“I'd start with fields both sides keep well: work email first, then name plus start date.”
Why it scores
Starts from fields that are reliable on both sides, not whatever columns exist.
- 2
Match exact keys first, then fuzzy matches with a confidence score
Say it like this
“Exact email matches go first, and that might cover 70%. Then fuzzy matching on cleaned-up names, so “Bob” finds “Robert” and swapped surnames still match, scored with date of birth and office as extra evidence.”
Why it scores
Exact before fuzzy, with real name problems named. Sounds hands-on.
- 3
Send low-confidence pairs to a person for review
Say it like this
“Anything under the threshold, maybe 90 pairs, goes to an HR partner as a sheet with both records side by side. I don't let a script guess on someone's payroll.”
Why it scores
Sends doubt to a person. A wrong merge in HR data means wrong pay.
- 4
Keep a mapping table and re-check it when either system changes
Say it like this
“The output is a mapping table with how and when each pair matched. It re-runs weekly, and new hires or changed emails get flagged.”
Why it scores
A mapping that stays current, not a one-time join.
StrongAll together
About a minute, out loud. The marked parts are the four steps; hover one to find it above.
Say a company bought a smaller one: 4,000 people in Workday, 600 in the acquired company's HR tool. I'd start with fields both sides keep well: work email first, then name plus start date.1 Exact email matches go first, and that might cover 70%. Then fuzzy matching on cleaned-up names, so “Bob” finds “Robert” and swapped surnames still match, scored with date of birth and office as extra evidence.2 Anything under the threshold, maybe 90 pairs, goes to an HR partner as a sheet with both records side by side. I don't let a script guess on someone's payroll.3 The output is a mapping table with how and when each pair matched. It re-runs weekly, and new hires or changed emails get flagged.4
WeakThe answer that fails
Sounds fine. Common. Scores low.
“I'd normalize the names first, lowercase them and trim the spaces, then run fuzzy matching with something like Levenshtein distance and merge anything above 85% similarity.”
Interviewer's note
Two John Smiths become one person. No review step, and nothing happens when either system changes.
PracticeTwo minutes, then they push back
Your turn
Two HR systems list the same people under different IDs. How do you match them?
Answer like you're in the room: about two minutes, no notes. Say it out loud, even alone. It feels different from thinking it. Then check your answer against what interviewers score.