One shared landing directory receives deliveries from many producers: clickstream exports, audit logs, invoice images, and order events, all mixed in the same path. Only the files named with the pattern orders-*.csv may ever reach the orders table; everything else must be ignored automatically, with no manual file handling between runs. Which ingestion approach meets this requirement?

1. A. A COPY INTO statement with a PATTERN clause of orders-*.csv, so only the files whose names match the pattern are loaded 💬 A
2. B. A staging load of every file in the directory, followed by a DELETE that removes the rows originating from non-order files 💬 B
3. C. A manual archive step that moves every non-order file to another prefix before each run, leaving only the matching files in place 💬 C
4. D. An Auto Loader stream with cloudFiles.pathGlobFilter set to orders-*.csv, so only files matching the glob pattern are ingested ✓ D

Reveal answer

Correct answer: An Auto Loader stream with cloudFiles.pathGlobFilter set to orders-*.csv, so only files matching the glob pattern are ingested

Reference: https://docs.databricks.com/aws/en/ingestion/cloud-object-storage/auto-loader/