FREE AWS Data Engineer DEA-C01 Exam Guide 2026: Pass First Try

AWS Certified Data Engineer - Associate (DEA-C01) Exam Guide 2026: The Only Walkthrough You Need

If your day job involves moving data into AWS, transforming it with Glue or EMR, modeling it in Redshift or Iceberg on S3, and keeping pipelines reliable with Step Functions and CloudWatch, then the AWS Certified Data Engineer - Associate (DEA-C01) is the single most relevant credential on your resume in 2026. It replaces the now-retired AWS Certified Data Analytics - Specialty (DAS-C01) as the primary AWS data credential, and it is explicitly positioned as a role-based Associate exam for data engineers, not data scientists or BI analysts.

This guide is written exclusively for the current 2026 exam version (DEA-C01). Every domain weight, fee, in-scope service, and policy below is cross-checked against the DEA-C01 Exam Guide PDF on docs.aws.amazon.com and the official exam page at aws.amazon.com/certification/certified-data-engineer-associate/. If a competitor guide still leans on DAS-C01 content (Kinesis Data Analytics for SQL, QuickSight deep dives, legacy EMR instance-fleet tuning), skip it - the emphasis on Glue, Redshift Serverless, Lake Formation, Iceberg, and orchestration is very different on DEA-C01.

DEA-C01 At-a-Glance (2026)

Sources: aws.amazon.com/certification/certified-data-engineer-associate/ and the DEA-C01 Exam Guide PDF (d1.awsstatic.com/training-and-certification/docs-data-engineer-associate/AWS-Certified-Data-Engineer-Associate_Exam-Guide.pdf).

---

Start Your FREE DEA-C01 Prep Today

Start FREE AWS Data Engineer Associate practice questionsPractice questions with detailed explanations

---

Who Should Sit DEA-C01

DEA-C01 is positioned as a role-based Associate exam for data engineers who design, build, secure, and operate data pipelines on AWS. You are a strong fit if:

Who should NOT start here? If you have never written a Glue job, never queried Athena on an S3 data lake, and never configured a Kinesis stream, DEA-C01 is too steep a first exam. Pass AWS Certified Cloud Practitioner (CLF-C02) or Solutions Architect Associate (SAA-C03) first and build 6-12 months of hands-on data pipeline experience before sitting DEA-C01.

Prerequisites and Recommended Background

AWS does not require any certification to sit DEA-C01. The official target-candidate description is:

• 2-3 years of experience in data engineering (understanding volume/variety/velocity, data modeling, performance and cost trade-offs, data pipelines)
• 1-2 years of hands-on AWS experience across core services
• Proficiency with at least one programming language (Python and SQL are assumed; Scala/Java helpful for Spark)
• Core data concepts: schema design, partitioning, indexing, batch vs streaming, CDC, OLTP vs OLAP, lakehouse/medallion patterns
• Familiarity with: Apache Spark, Apache Hadoop ecosystem basics, open table formats (Iceberg/Hudi/Delta), orchestration (Airflow/Step Functions), IAM

If any of those are gaps, close them before buying a voucher. Reading the AWS Glue Developer Guide + Amazon Redshift Database Developer Guide front-to-back is 3-4 weekends and pays back enormously during the exam.

---

Build DEA-C01 Mastery with FREE Practice Questions

Access FREE AWS Data Engineer Associate practice questionsPractice questions with detailed explanations

---

DEA-C01 vs DAS-C01 vs Microsoft DP-700: Positioning in 2026

DEA-C01 sits at the intersection of three roles that were historically split. Understanding where it fits vs competing credentials sharpens your study focus:

Bottom line: DEA-C01 is the 2026 AWS data credential. If you hold DAS-C01 and it is nearing expiry, DEA-C01 is the recommended upgrade path. DP-700 is only relevant if your org runs Microsoft Fabric - they are not substitutes.

---

DEA-C01 Domain Blueprint (2026 Weights - Verified)

The official Exam Guide breaks DEA-C01 into four content domains with fixed percentage weights. Always verify current weights from the latest DEA-C01 Exam Guide PDF before you schedule, as AWS refreshes content periodically.

Source: AWS Certified Data Engineer - Associate (DEA-C01) Exam Guide - Content Outline.

Three observations experienced candidates wish they had known earlier:

1. Ingestion + Transformation is 34% - by far the largest single domain. Kinesis family, MSK, DMS, Glue (all flavors), EMR Serverless, AppFlow, and Lambda pipelines all live here. If you learn Glue and Kinesis cold, you have already earned ~25% of the exam.
2. Store Management (26%) is NOT just S3. Expect scenario questions comparing Redshift RA3 vs Serverless, Aurora vs DynamoDB for OLTP feeds, Iceberg vs Hudi vs Delta for transactional lakes, and DocumentDB/Neptune/Timestream/QLDB for specialty workloads.
3. Security and Governance (18%) punches above its weight. Lake Formation grants (column/row/cell-level), KMS envelope encryption, VPC endpoints for data traffic, and Macie for PII detection are heavily tested - and they are exactly the topics new data engineers skip.

Domain 1: Data Ingestion and Transformation (34%)

This is the biggest domain and covers the full pipeline from source to lake.

Ingestion services in scope:

• Amazon Kinesis Data Streams - shard math (1 MB/s or 1,000 rec/s per shard write; 2 MB/s per shard read / or dedicated 2 MB/s per consumer with Enhanced Fan-Out), provisioned vs on-demand, 24h-365d retention, ordering by PartitionKey.
• Amazon Kinesis Data Firehose (now branded Amazon Data Firehose) - near real-time delivery to S3/Redshift/OpenSearch/Splunk, buffering (60s-900s or 1-128MB), format conversion to Parquet/ORC, dynamic partitioning, built-in Lambda transformation.
• Kinesis Video Streams - ingestion of video/audio; less weight but appears.
• Amazon MSK / MSK Serverless - managed Kafka; IAM auth, SASL/SCRAM, TLS; MSK Connect for Kafka Connect workloads.
• AWS DMS - homogeneous + heterogeneous migrations, CDC via transaction logs, task types (full load, CDC, full+CDC), replication instance sizing, endpoints.
• AWS Glue streaming ETL - continuous Spark Structured Streaming jobs reading from Kinesis/MSK.
• Amazon AppFlow - SaaS-to-AWS flows (Salesforce, SAP, ServiceNow, Google Analytics, etc.).
• AWS Transfer Family - managed SFTP/FTPS/FTP/AS2 to S3 or EFS.
• AWS Lambda - small-batch HTTP ingestion and event-driven pipelines.
• Amazon SQS / SNS / EventBridge - decoupling and fan-out patterns.

Transformation services in scope:

• AWS Glue Studio + Glue ETL + Glue DataBrew + Glue Flex + Glue Streaming - visual ETL, Spark-based Python/Scala jobs, low-code DataBrew recipes, Flex execution class for cost, crawlers, bookmarks for incremental processing.
• Amazon EMR (on EC2) + EMR Serverless + EMR on EKS - Spark, Hive, Presto/Trino, Hudi, Iceberg runtime. When to pick EMR over Glue: long-running clusters, Hadoop ecosystem tools beyond Spark, large (>100 DPU) sustained workloads.
• AWS Lambda for lightweight transforms (15-minute max, 10 GB memory).
• Amazon Athena (including Athena for Apache Spark) - federated queries, CTAS, views, Iceberg DDL, workgroups, result reuse.
• Amazon Redshift + Redshift ML + Redshift Serverless + Redshift Spectrum - SQL transforms on columnar store; Spectrum for querying S3 directly.

Highest-yield sub-topics:

• Glue vs EMR decision: pick Glue for serverless Spark, AWS-native integrations, short-lived ETL; pick EMR when you need Hive/Presto/Flink/HBase/long clusters or custom Spark versions.
• Kinesis shard math: given X MB/s ingest and Y KB record size, compute shard count. On-demand scales automatically; provisioned requires manual SplitShard/MergeShards.
• Firehose vs Streams: Firehose = near-real-time (~60s+), no custom consumer, managed delivery. Streams = sub-second, custom consumers, shard control.
• Glue crawlers + Data Catalog: automatic schema inference, partition discovery, versioned tables, classifiers for custom formats.
• Glue bookmarks: track processed state across runs for incremental ETL (enable on the job).
• DMS CDC: transactional log reads, ongoing replication, validation mode for row-level checks.

Domain 2: Data Store Management (26%)

The second-largest domain covers both lake and warehouse patterns.

Object storage + data lake:

• Amazon S3 - storage classes (Standard, Intelligent-Tiering, Standard-IA, One Zone-IA, Glacier Instant Retrieval, Glacier Flexible Retrieval, Glacier Deep Archive), lifecycle policies, versioning, Object Lock (compliance vs governance), event notifications, multipart upload, strong read-after-write consistency.
• S3 Intelligent-Tiering - auto-moves objects between frequent/infrequent/archive tiers based on access patterns.
• Glacier tiers - Instant Retrieval (ms), Flexible Retrieval (minutes-hours), Deep Archive (12h). Minimum storage durations matter for cost questions.
• AWS Lake Formation - centralized data lake permissions, LF-tags, column/row/cell-level filters, cross-account data sharing, named resources vs tag-based access.
• AWS Glue Data Catalog - Hive-compatible metastore; tables, partitions, databases; resource links for cross-account.
• Open table formats: Apache Iceberg (default recommendation in AWS docs for new tables - ACID, time travel, schema evolution, hidden partitioning); Apache Hudi (CDC-optimized, copy-on-write vs merge-on-read); Delta Lake (Databricks-origin; supported in Glue/EMR).

Warehouses and OLTP:

• Amazon Redshift RA3 + Redshift Serverless + AQUA - RA3 nodes (ra3.xlplus/4xl/16xl) separate compute and storage (RMS), Serverless auto-scales with RPUs, AQUA accelerates scans via FPGAs.
• Distribution styles: KEY (co-locates matching values), EVEN (round-robin), ALL (small dimension tables), AUTO (Redshift decides).
• Sort keys: COMPOUND (default, filters on leading columns) vs INTERLEAVED (equal weight across columns; rarely recommended in 2026).
• Amazon Aurora (MySQL/Postgres compatible) - serverless v2 for variable load; zero-ETL to Redshift for analytics.
• Amazon DynamoDB - OLTP key-value; single-digit ms latency; Streams for CDC into Lambda; global tables for multi-Region.
• Amazon Neptune - graph DB (property graph + RDF).
• Amazon Timestream - time-series DB with tiered storage (memory + magnetic).
• Amazon DocumentDB - MongoDB-compatible.
• Amazon QLDB - immutable, cryptographically verifiable ledger (note: QLDB was announced for end-of-support; verify current status in the exam guide).

Highest-yield sub-topics:

• Redshift distribution + sort key selection for a given join/filter pattern.
• S3 lifecycle transitions and expiration with minimum-duration cost implications.
• Iceberg vs Hudi vs Delta trade-offs - default to Iceberg in AWS-native pipelines.
• Lake Formation column-level security vs IAM alone (LF adds fine-grained + tag-based).
• Zero-ETL Aurora -> Redshift and DynamoDB -> Redshift as CDC alternatives.

Domain 3: Data Operations and Support (22%)

This domain tests whether your pipelines stay running and are debuggable.

Orchestration:

• AWS Step Functions - Standard (up to 1 year, per-transition pricing) vs Express (up to 5 min, per-invocation pricing); Map state for parallel iteration; distributed Map for large-scale S3 fan-out; Retry/Catch blocks; callback tokens for human-in-the-loop.
• Amazon MWAA (Managed Workflows for Apache Airflow) - managed Airflow 2.x clusters; DAG authoring in Python; choose MWAA over Step Functions when you need Python-rich DAGs, existing Airflow code, or multi-cloud orchestration.
• Amazon EventBridge Scheduler + Rules + Pipes - cron/rate scheduling, pattern-based routing, point-to-point source-to-target with enrichment.

Monitoring + observability:

• Amazon CloudWatch Metrics (standard 1-min vs high-resolution 1-sec), Logs (log groups/streams/retention, metric filters, Logs Insights), Alarms (threshold/composite), CloudWatch Dashboards.
• AWS X-Ray - distributed tracing for serverless pipelines (Lambda, Step Functions).
• Amazon Managed Service for Prometheus + Grafana - for open-source monitoring stacks.

Error handling:

• Dead-letter queues (SQS/SNS) for Lambda and Kinesis consumers.
• Step Functions Retry/Catch patterns and exponential backoff.
• Glue job bookmarks + retries with max-retry config.
• Kinesis Data Streams checkpointing (via KCL DynamoDB lease table).

Data quality:

• AWS Glue Data Quality - built-in rules engine (powered by Deequ under the hood); rulesets; evaluation during jobs or on-demand; publish results to CloudWatch.
• AWS Glue DataBrew data quality profiles - statistical profiling + anomaly detection.
• Amazon Deequ (open source) - data unit tests; available as library on EMR.

Highest-yield sub-topics:

• Step Functions vs MWAA decision: Step Functions for AWS-native event-driven flows; MWAA when you already have Airflow DAGs or complex Python logic.
• CloudWatch Logs Insights queries: fields, filter, stats, sort, parse.
• Glue job retries + max concurrent runs - preventing duplicate processing on retry.
• DLQ patterns - when to DLQ at the source (Kinesis) vs at the consumer (SQS tied to Lambda).
• Glue Data Quality rulesets - Completeness, Uniqueness, ColumnValues, RowCount, StandardDeviation rule types.

Domain 4: Data Security and Governance (18%)

Smallest domain but disproportionately represented in the tougher questions.

Identity and access:

• AWS IAM - policies (identity-based, resource-based, permission boundaries, session, SCP), evaluation logic (explicit Deny > explicit Allow > implicit Deny), condition keys, roles vs users.
• AWS Lake Formation grants - GRANT SELECT at database/table/column/row/cell level, LF-tags for tag-based access control (TBAC), data filters, cross-account sharing via AWS RAM.
• Resource-based policies on S3, KMS, Glue Catalog, Lambda, SNS, SQS.

Encryption:

• AWS KMS - customer managed (CMK) vs AWS managed vs AWS owned; key policies + grants; automatic rotation (365 days); envelope encryption + GenerateDataKey; multi-Region keys.
• S3 encryption - SSE-S3, SSE-KMS, SSE-C, DSSE-KMS (dual-layer), bucket keys for KMS cost reduction. Default encryption is enabled on all new buckets since 2023 - but older buckets may still be opt-in.
• Redshift encryption - at rest (KMS) + in transit (SSL required option).
• EBS + RDS + Aurora + DynamoDB encryption at rest.

Network isolation:

• VPC endpoints (Gateway for S3/DynamoDB; Interface/PrivateLink for most others) to avoid traversing the public internet.
• S3 bucket policies + access points for multi-tenant access patterns.
• Security groups + NACLs on Redshift, MSK, EMR clusters.

Audit + compliance:

• AWS CloudTrail - management + data events; CloudTrail Lake for queryable history.
• Amazon Macie - ML-based PII/sensitive data discovery on S3.
• AWS Config - resource configuration history and compliance rules.
• Amazon GuardDuty - threat detection across CloudTrail/VPC Flow/DNS logs.

Highest-yield sub-topics:

• Lake Formation vs IAM - when do you need both? LF adds fine-grained (column/row/cell) on top of IAM.
• KMS envelope encryption - why GenerateDataKey + local encryption beats calling Encrypt directly on large objects.
• Cross-account data sharing - Lake Formation via RAM; Redshift data sharing via datashares.
• Macie findings - sensitive data identifiers, custom data identifiers, managed jobs vs one-time.
• Bucket policy conditions - aws:SourceVpce, aws:SecureTransport, aws:PrincipalOrgID.

---

Cost & Registration (2026)

How to register: Create an AWS Certification account at aws.amazon.com/certification/, then schedule with Pearson VUE (testing center or OnVUE online proctor). Government-issued photo ID must match your Certification profile name exactly. Book 2-3 weeks out to lock your preferred slot, longer for weekend test centers in major metros.

Recertification Strategy (2026)

AWS Associate certifications are valid for 3 years. You can recertify two ways:

1. Pass DEA-C01 again at its latest series code.
2. Pass a higher-tier AWS certification in the intervening 3 years - the new cert automatically renews any Associate-level certifications you hold. Common upgrades for DEA-C01 holders:
   • AWS Certified Solutions Architect - Professional (SAP-C02) - if you also do architecture.
   • AWS Certified DevOps Engineer - Professional (DOP-C02) - if you own CI/CD for data platforms.
   • AWS Certified Machine Learning Engineer - Associate or AWS Certified Machine Learning - Specialty (MLS-C01) - natural extension for ML pipelines.
   • AWS Certified Security - Specialty (SCS-C02) - for data teams in regulated industries.

AWS does not offer a CEU/continuing-education credit system. Recertification is pass-an-exam only.

10-12 Week DEA-C01 Study Plan (For Working Data Engineers)

This plan assumes 8-10 hours/week and the recommended ~2 years of data engineering experience. Cut to 6-8 weeks if full-time student with more hours; extend to 14-16 weeks if coming from a non-data engineering background.

Time Allocation (Match the Blueprint, Not Your Comfort Zone)

Recommended Resources (FREE-First, Then Paid)

Hands-On Labs Are Non-Negotiable

DEA-C01 tests implementation knowledge. Theory alone will not get you to 720. Build these six end-to-end labs in an AWS Free Tier + minimal-paid account before test day:

1. Partitioned S3 data lake with Glue + Athena: crawl CSV/JSON, convert to Parquet with Glue ETL, query with Athena, add partition projection.
2. Streaming pipeline: Kinesis Data Streams -> Firehose (Parquet + dynamic partitioning) -> S3 -> Athena.
3. Redshift data warehouse: load via COPY from S3, pick KEY/EVEN/ALL distribution appropriately, create materialized views, query S3 via Spectrum.
4. CDC pipeline: DMS from RDS Postgres to S3 (Parquet) + Aurora zero-ETL to Redshift.
5. Orchestrated workflow: Step Functions Standard that runs Glue crawler -> Glue ETL -> Redshift COPY -> Glue Data Quality check, with Retry/Catch and an SNS DLQ.
6. Secured lake: Lake Formation column-level permissions + LF-tags + KMS-encrypted S3 + VPC endpoints; scan with Macie and audit with CloudTrail.

If you have built all six, the exam's scenario items will read like your own system.

Test-Day Strategy

Before the exam:

• Confirm the name on your AWS Certification profile matches your government photo ID exactly.
• If online-proctored (OnVUE), run the Pearson VUE system test 24+ hours in advance; webcam, microphone, and bandwidth failures cost slots.
• Clear your workspace: no notes, no secondary monitors, no phone within reach. Proctor will require a 360-degree room scan.
• Sleep. Caffeine is fine if routine, skip if not.

During the exam:

• 130 minutes / 65 questions = exactly 2 minutes per question. Flag and move on if you spend 2.5 minutes.
• Read the last sentence first. Case-study style questions hide the actual ask in the last line ("most cost-effective," "least operational overhead," "highest throughput").
• Eliminate first. On 4-option questions you can usually discard 2 obviously wrong answers; pick the AWS-idiomatic one between the final two. Idiomatic = managed service, least code, least ops overhead, most loosely coupled, native integration over custom plumbing.
• Multi-response questions tell you how many to choose ("Select TWO," "Select THREE"). No partial credit on DEA-C01.
• Familiarize yourself with service short-names beforehand - MSK, MWAA, DMS, AppFlow, KCL, KPL - so the stem does not surprise you.

After the exam:

• Provisional pass/fail appears on screen at the test center or at OnVUE session close.
• Official result + scaled score (100-1000; pass = 720) arrives within 5 business days via email.
• Skill-area breakdown shows "Meets competency" / "Needs improvement" per domain.
• Digital badge is Credly-issued and shareable on LinkedIn within a day of passing.

Common Pitfalls That Sink First-Time Scores

1. Glue vs EMR default. Candidates default to Glue on every question. Know when EMR wins: long-running clusters, non-Spark Hadoop tools (Hive, Presto, HBase, Flink), sustained heavy workloads, custom Spark versions, on-EC2 reserved-instance economics.
2. Kinesis shard math errors. Memorize: 1 MB/s or 1,000 rec/s write per shard; 2 MB/s read per shard (shared) or 2 MB/s per consumer with Enhanced Fan-Out. Questions will give you throughput and record size - compute shards.
3. Redshift distribution style confusion. KEY = co-locates on a join key (pick when two big tables join). EVEN = round-robin (pick when no natural key). ALL = copy to every node (only for small dimensions, <~3M rows). AUTO = let Redshift choose.
4. IAM vs Lake Formation permissions. IAM alone cannot give column-level access. Lake Formation layers on top; both must allow the action. Know "Use only IAM access control" vs Lake Formation modes.
5. Confusing Firehose and Data Streams. Firehose = managed delivery, ~60s+ latency, no custom consumers. Streams = sub-second, custom consumers, shards you manage (or on-demand).
6. Missing Step Functions Standard vs Express. Standard = up to 1 year, priced per transition. Express = up to 5 min, priced per invocation + duration. Express for high-volume short workflows (API backends); Standard for orchestrations measured in minutes-hours.
7. Over-engineering with MWAA. Unless the question mentions existing Airflow DAGs or Python-heavy orchestration, Step Functions is typically the AWS-idiomatic answer.
8. Weak on Iceberg/Hudi/Delta. Default to Iceberg for new AWS-native transactional lakes; Hudi for CDC-heavy upsert workloads; Delta when Databricks interop matters.
9. Glue bookmarks left disabled. For incremental ETL questions, enable bookmarks on the job and use a job-specific state.
10. Trusting DAS-C01 content. If a resource spends chapters on Kinesis Data Analytics for SQL or QuickSight dashboard design, it is the wrong prep. DEA-C01 is engineering, not analytics.

Career Value: AWS Data Engineer Salary in 2026

AWS-certified data engineers consistently rank in the top tier of cloud IT pay across every major salary survey (Global Knowledge, Skillsoft, Dice, PayScale, Built In).

Career Ladder

Intangibles that move the needle: AWS re:Invent ANT/DAT track networking, open-source contributions to Iceberg/Hudi/Deequ/Glue, public post-mortems of your own data incidents, and a portfolio of reusable Glue/Spark/Step Function templates.

How DEA-C01 Fits Into the Broader AWS Certification Path

Recommended 2026 data-engineer path: CLF-C02 (optional) -> SAA-C03 or DVA-C02 -> DEA-C01 -> SAP-C02 or MLS-C01.

Frequently Missed 2026 Details (That Competitor Guides Get Wrong)

• DEA-C01 GA date was March 12, 2024, following a beta from November 2023 to January 2024. Any "DAS-C01 replacement beta tips" post is out of date.
• Firehose is now branded "Amazon Data Firehose" (no longer "Kinesis Data Firehose") since early 2024.
• Kinesis Data Analytics for SQL was deprecated; the current managed streaming analytics service is Amazon Managed Service for Apache Flink.
• Aurora zero-ETL to Redshift (GA 2023) and DynamoDB zero-ETL to Redshift (GA 2024) are in scope as CDC alternatives.
• EMR Serverless (GA 2022) is the default-recommended EMR mode for most new workloads - classic EMR on EC2 is tested but usually not the "right" answer.
• Iceberg is a first-class citizen in Glue, Athena, EMR, and Redshift Spectrum. Hudi and Delta are supported but Iceberg is AWS-recommended for new tables.
• Athena for Apache Spark (GA 2022) brings notebook-style Spark sessions inside Athena workgroups.
• Redshift ML in-scope for SQL-based model training and inference.
• Glue Data Quality (GA 2023) is part of the Operations & Support blueprint.
• AWS Glue Flex execution class cuts cost ~34% for non-SLA batch jobs.

---

Keep Training with FREE DEA-C01 Practice

Practice AWS Data Engineer Associate scenario questionsPractice questions with detailed explanations

---

Official Sources Used

• AWS Certified Data Engineer - Associate exam page (aws.amazon.com/certification/certified-data-engineer-associate/)
• AWS Certified Data Engineer - Associate (DEA-C01) Exam Guide PDF (d1.awsstatic.com/training-and-certification/docs-data-engineer-associate/AWS-Certified-Data-Engineer-Associate_Exam-Guide.pdf)
• AWS Training and Certification blog - DEA-C01 launch posts + "Demystifying your AWS Certification exam score"
• AWS Recertification Policy (aws.amazon.com/certification/policies/recertification/)
• AWS Service Developer Guides (Glue, Redshift, Kinesis, S3, Lake Formation, Athena, EMR, Step Functions, DMS, MSK, MWAA)
• Glassdoor, Built In, PayScale, ZipRecruiter, Dice 2026 salary references

Certification details, fees, and blueprint weights may be revised by AWS. Always confirm current requirements directly on aws.amazon.com/certification/certified-data-engineer-associate before scheduling.