Exams / DEA-C01 / Study guide / Schema design, conversion and lineage

Schema design, conversion and lineage: DEA-C01 study notes and practice questions

What to know about Schema design, conversion and lineage for the DEA-C01 exam, from the Data Store Management domain (26% of the exam guide), then 1 free question to test it. Part of the DEA-C01 study guide.

Last verified October 2026 against current AWS documentation ·How we write our questions

What to know for the exam

Schema evolution questions describe a source that adds or changes fields and ask how to keep pipelines and queries working.

Typical changes

Adding a column: safe in most formats; old files return nulls for it.
Renaming or reordering: breaks positional formats like CSV and can confuse Parquet tables read by name or index.
Changing a type (int to string): often breaks readers unless the format supports promotion.
Removing a column: queries referencing it fail.

Tools that handle change

Tool	How it helps
Apache Iceberg	Add, drop, rename, reorder and widen types as metadata changes, with no rewrite and no broken queries
Glue crawlers	Schema change policy: update the table, add new columns only, or log and ignore
Glue DynamicFrames	Tolerate inconsistent types with choice types; `resolveChoice` casts or splits them
Glue jobs	`enableUpdateCatalog` with `updateBehavior` to add new columns as they write
Glue Schema Registry	Validates producer schemas for Kinesis, MSK and Kafka with compatibility modes (BACKWARD, FORWARD, FULL)
Avro / Parquet	Store the schema with the data; Avro has rules for defaults and evolution
Redshift SUPER	Stores semi-structured JSON whose fields change, queried with PartiQL
DMS	Can apply source DDL changes to targets

Practices

Keep a raw zone with the original data so you can reprocess after a schema fix.
Choose BACKWARD compatibility so new consumers can read old data.
Version schemas and test changes in CI.
Alert on schema drift (EventBridge events from the Data Catalog or crawler logs).

Common traps

A crawler overwriting a table schema that downstream jobs rely on. Use “add new columns only” or log changes.
CSV with reordered columns silently loading values into the wrong columns.
Rewriting a huge table to add one column when Iceberg could do it in metadata.

Try the 1 practice question

Practice Schema design, conversion and lineage Take the free practice exam

Question 1

Harder ·Link to this question

A healthcare-analytics company plans to migrate an Oracle data warehouse to Amazon Redshift. The schema has about 600 tables, 80 views and 150 stored procedures written in PL/SQL. Some views and procedures use Oracle-only functions.

The company must know which objects convert automatically, which need manual work, and how much effort the manual work will take, before it commits to a migration plan.

Which step should the team take first?

Choose one answer.

A. Create an AWS DMS replication task with the full-load option, and run it to see which tables fail to load into Redshift.
B. Run an AWS Glue crawler on the Oracle database, and review the tables it discovers in the Data Catalog for the conversion plan.
C. Rewrite all 150 procedures in Python on AWS Lambda first, and then move the data with DMS once the procedures run correctly.
D. Use AWS DMS Schema Conversion to assess the Oracle schema, convert what it can, and report the objects needing manual work.

Show the answer and explanation

Answer: D

A. Wrong: A DMS replication task moves rows, and it does not convert the schema objects such as procedures and views. The option reveals data load problems but not the conversion effort.
B. Wrong: A crawler reads table metadata into the catalog, but it does not analyze procedures or views for conversion. The option gives an incomplete picture of the work.
C. Wrong: Rewriting every procedure before knowing which ones convert automatically commits the team to the most effort before the assessment. The option skips the step that gives the plan its scope.
D. Right: Correct. DMS Schema Conversion assesses the source schema against the Redshift target and converts the objects it supports. Its assessment report lists the objects that need manual changes, which gives the team the effort estimate it needs. Learn more: Converting database schemas using DMS Schema Conversion (opens in a new tab)

Explanation

Before a heterogeneous migration, the team needs to know which schema objects convert automatically and how much manual work remains. DMS Schema Conversion assesses the Oracle schema against Redshift, converts what it can, and reports the objects that need manual changes. Replication, catalog crawling and early rewrites do not produce that assessment.

Study

Want more on Schema design, conversion and lineage? The 5 full-length practice sets have 6 more Schema design, conversion and lineage questions, 325 in all, none of them in the free exam. One payment of $14.99, no subscription. See the practice sets

Study Schema design, conversion and lineage

AWS DMS Schema Conversion