You are responsible for a Snowflake data pipeline that loads data from various external sources into a data warehouse. You have implemented Snowpipe for continuous data ingestion from AWS S3. One of your data sources is known to occasionally produce corrupted data files. You need to design a resilient pipeline that can handle these corrupted files without halting the entire data loading process and allows for the identification and quarantine of these files for further investigation. Assume you have a table called 'RAW DATA' and you are using a file format 'MY CSV FORMAT. Select TWO actions which best address this situation:
Correct Answer: B,C
Option B directly addresses the requirement for a resilient pipeline by utilizing 'ERROR INTEGRATION' to log errors and quarantine corrupt files. Setting ON_ERROR to CONTINUE avoids pipeline interruption, and creating a task to query the error stage and move files to a quarantine bucket achieves the desired outcome. Option C would be the 2nd best option because it provides for pre-processing and validation, which avoids loading the corrupted records into the RAW_DATA table in the first place. Option A only skips the file and doesn't log the errors or quarantine the file. Option D creates a stream on the RAW DATA table, but we would like to avoid writing erroneous records to the RAW DATA table in the first place. 'VALIDATE only checks the existing data, so it would not solve any problems in Data Load.