Finding Duplicate Records in a Large Unstructured Dataset with AWS Glue FindMatches

Why the Answer Is Correct

AWS Glue FindMatches is a managed transform that uses machine learning to identify duplicate or matching records in a dataset that already exists, and it needs no labeled training data. Because it is a single managed transform inside Glue, the company adds duplicate detection with essentially no code, which is exactly what the least-code-development requirement asks for. The votes were unanimous at 80 for D, and Saransundar and feelgoodfactor both highlight that FindMatches is purpose-built for duplicates and scales to large datasets.

Why the Other Options Are Wrong

Amazon SageMaker Data Wrangler (C) does offer a Drop duplicates transform, as conrad2023 noted, but that is a row-level exact-duplicate cleanup inside a broader interactive data preparation tool, not fuzzy record matching across several key attributes on a large unstructured dataset, and it pulls in the overhead of provisioning a Data Wrangler session. Amazon QuickSight ML Insights (B) is a business-intelligence visualization feature for explaining anomalies in dashboards; using it to build a custom deduplication model would mean writing that model yourself, which is the opposite of least code. Amazon Mechanical Turk (A) is a crowdsourcing service that would require paying people to compare records, which is manual, expensive, and not code-based at all.

Community Comment Notes

The community was decisive here, with all 80 votes for D. GiorgioGss linked the AWS Glue launch announcement describing FindMatches as letting you identify duplicate or matching records in your dataset, and Saransundar emphasized that it works without labeled training data. The only dissenting comment, from nakidal495, expressed uncertainty rather than a substantive counterargument.