DEA-C01 Practice Question 58: DMS Data Validation
AWS Certified Data Engineer – Associate (DEA-C01) · Test 6 · question 8 of 10
An AWS DMS task is migrating an Oracle database to Amazon Aurora PostgreSQL with full load plus CDC. Before the cutover, the project lead must be able to demonstrate, row by row, that the target matches the source, and must be given the primary keys of any rows that do not match so they can be investigated.
Which approach provides this?
Show answer and explanation
Answer:
DMS has this built in: "AWS DMS provides support for data validation to ensure that your data was migrated accurately from the source to the target. If enabled, validation begins immediately after a full load is performed for a table. Validation compares the incremental changes for a CDC-enabled task as they occur." The comparison is genuinely row by row: "AWS DMS compares each row in the source with its corresponding row at the target, verifies the rows contain the same data, and reports any mismatches." Per-table statistics (ValidationState, ValidationPending, ValidationFailed) give the overall picture, and the detail the lead needs is in a control table: "During validation, AWS DMS creates a new table at the target endpoint: awsdms_control.awsdms_validation_failures_v1… You can query this table to help troubleshoot validation errors", with a KEY column holding the primary key and a DETAILS column listing the mismatched column values. (Validation requires a primary key or unique index on each table.)
Option A compares totals, not content. Two tables can have identical row counts while columns differ, characters were substituted, or numeric precision was lost in the Oracle-to-PostgreSQL type mapping, and a count gives no keys to investigate.
Option B reports what DMS failed to apply. Rows that applied successfully but arrived with altered values never appear in the task log, which is exactly the class of problem a pre-cutover sign-off needs to rule out.
Option D checks the target against rules you write about the target. Glue Data Quality can assert completeness or ranges, but it never reads the Oracle source, so it cannot prove the two systems agree.
Reference: AWS DMS data validation