DEA-C01 Practice Question 81: FindMatches Duplicate Customer Records
AWS Certified Data Engineer – Associate (DEA-C01) · Test 9 · question 1 of 10
A retailer merges customer tables from two acquired companies. The same person often appears in both tables with different name spellings, partial addresses and missing fields, and there is no shared customer ID. The data engineering team wants to identify duplicate customers inside an AWS Glue ETL job without writing matching rules by hand and without ML expertise. Which approach should they use?
Show answer and explanation
Answer:
The FindMatches transform *"enables you to identify duplicate or matching records in your dataset, even when the records do not have a common unique identifier and no fields match exactly,"* and it *"will not require writing any code or knowing how machine learning works."* You teach it by labeling example records, and then call the transform from a Spark-based AWS Glue job. The output can add a match_id column or remove duplicate rows (option D).
- A is wrong because DISTINCT removes only rows that are identical in every column. Records with different spellings or missing fields would all remain. - B is wrong because hand-written rules are brittle and time-consuming, and the team wants to avoid writing matching logic. - C is wrong because a crawler infers schemas and populates the Data Catalog. It does not compare records or find duplicates.
Reference: Record matching with AWS Lake Formation FindMatches