A company is reading data from various customer databases that run on Amazon RDS. The databases contain many inconsistent fields. For example, a customer record field that is place_id in one database is location_id in another database. The company wants to link customer records across different databases, even when many customer record fields do not match exactly.
Which solution will meet these requirements with the LEAST operational overhead?

A. Create an Amazon EMR cluster to process and analyze data in the databases. Connect to the Apache Zeppelin notebook, and use the FindMatches transform to find duplicate records in the data.
B. Create an AWS Give crawler to crawl the databases. Use the FindMatches transform to find duplicate records in the data. Evaluate and tune the transform by evaluating performance and results of finding matches.
C. Create an AWS Glue crawler to crawl the data in the databases. Use Amazon SageMaker to construct Apache Spark ML pipelines to find duplicate records in the data.
D. Create an Amazon EMR cluster to process and analyze data in the databases. Connect to the Apache Zeppelin notebook, and use Apache Spark ML to find duplicate records in the data. Evaluate and tune the model by evaluating performance and results of finding duplicates.