Databricks DE · Topic 12.1

Map Skills to the Exam Domains

The DE Associate exam is 7 weighted domains. Study to the weights — map each domain to the stages you've already built, and rehearse the kind of code each domain tests.

7 domains45 Q · 90 min$200 · 2 yrsstudy to weights

The one-liner

The exam scores 7 weighted domains; Ingestion (21%) + Transformation (22%) + Jobs (16%) ≈ 59%. Map domains → your Stages 1–11 and allocate study time by weight. Verify current weights on the official guide.
📋 Analogy: the blueprint is a grading rubric leaked in advance — you know which sections carry the most points, so you study to the weights instead of cramming everything equally.

Domains by weight — click to see which stage covers it

▸ bigger bar = more of your exam score

The 7 domains — weight, coverage & an exam-style snippet

1 · Databricks Intelligence Platform · 6% · Stage 1

Lakehouse, workspace/architecture, medallion. Lowest weight — know the concepts.

-- Recognize the 3-level Unity Catalog namespace (a platform fundamental):
SELECT * FROM main.sales.orders;   -- catalog.schema.table

2 · Data Ingestion and Loading · 21% · Stage 4

Auto Loader (cloudFiles), COPY INTO, Lakeflow Connect, schema evolution.

# Auto Loader — incremental file ingestion (high-yield):
(spark.readStream.format("cloudFiles")
   .option("cloudFiles.format", "json")
   .option("cloudFiles.schemaLocation", "/Volumes/main/raw/_schema")
   .load("/Volumes/main/raw/orders")
   .writeStream.option("checkpointLocation", "/Volumes/main/raw/_ckpt")
   .toTable("main.bronze.orders"))

3 · Data Transformation and Modeling · 22% · Stages 2, 5, 6

Highest weight. Delta, MERGE, SDP, expectations, PySpark/Spark-SQL ETL.

-- MERGE (upsert) — a classic exam item:
MERGE INTO main.silver.customers t
USING updates s ON t.id = s.id
WHEN MATCHED THEN UPDATE SET *
WHEN NOT MATCHED THEN INSERT *;

4 · Working with Lakeflow Jobs · 16% · Stage 7

Tasks & dependencies, scheduling with Quartz cron, retries, alerts.

# 6-field Quartz cron (NOT 5-field Unix) — frequently tested:
schedule: { quartz_cron_expression: "0 0 2 * * ?", timezone_id: "UTC" }  # 02:00 daily

5 · Implementing CI/CD · 10% · Stage 11

Databricks Asset Bundles, Git folders, dev/prod targets, deploy flow.

databricks bundle validate          # check config (CI on PRs)
databricks bundle deploy -t prod    # promote on merge

6 · Troubleshooting, Monitoring & Optimization · 10% · Stages 3, 7, 10

OPTIMIZE/ZORDER/Liquid Clustering, VACUUM, query profile, run monitoring.

-- Liquid Clustering — added to the current syllabus (know it):
ALTER TABLE main.silver.events CLUSTER BY (event_date, region);
OPTIMIZE main.silver.events;

7 · Governance and Security · 15% · Stages 8, 9

Unity Catalog grants, row filters / column masks, ABAC, Delta Sharing.

-- A GRANT on the UC 3-level namespace (high-yield governance item):
GRANT SELECT ON TABLE main.sales.orders TO `analysts`;

Study allocation (by weight)

▸ proportional time blocks — spend where the points are

Ingestion + Transformation + Jobs ≈ 59% → the bulk of your time.

Logistics (verified — re-check before booking)

Why it matters

🗓️ Real-world use case: two weeks out — most time on Stages 2/4/5 (Delta + ingestion + pipelines), a solid block on Stage 8 (governance, 15%), a quick review of Stage 1 (6%).

Uses, edge cases & limitations

Common gotchas