Databricks interview questions
Databricks interview questions on control plane vs data plane, compute types, Unity Catalog, Delta Lake, medallion, DLT, and Photon.
210 questions
Beginner
Control plane vs data plane, Delta basics, and bronze / silver / gold.
Senior
Unity Catalog grants, MERGE file rewrites, OPTIMIZE, Photon fallbacks.
Scenarios
Workspace vs VPC, job-cluster bills, small-file MERGE, DLT expectations.
By topic
Each topic has the lesson and the simulator alongside the questions.
Databricks Architecture
Control plane vs data plane: workspace in Databricks SaaS, clusters and files in your cloud.
Compute types
All-purpose clusters, job clusters that terminate, and SQL warehouses for BI.
SQL Warehouses
Serverless vs Pro vs Classic, the caches behind a fast dashboard, and why size and cluster count are different dials.
Unity Catalog
Three-level namespace, governed storage, and how permissions attach to catalogs, schemas, and tables.
Delta Lake
ACID transactions on object storage: the _delta_log, Parquet data files, time travel, and OPTIMIZE.
Time Travel & VACUUM
Read an old version from the commit log, RESTORE as a new forward commit, and the VACUUM that ends time travel.
Medallion Architecture
Bronze keeps the raw files. Silver is typed and joined. Gold is the metric table BI queries.
Delta MERGE
Upserts a change set into Delta by rewriting only the files the join key touches.
OPTIMIZE & Clustering
Compact tiny files, Z-ORDER for skipping, and liquid clustering on write.
Liquid Clustering
CLUSTER BY replaces rigid directories and full Z-ORDER rewrites — and lets you change the keys later.
Auto Loader
Incremental ingest from object storage with notifications, checkpoint, and schema evolution.
Delta Live Tables
Declare medallion tables and expectations. DLT infers the graph and stops downstream on failure.
Workflows
A job is a DAG of tasks with real parallelism, per-task retries, and a repair run that skips what already passed.
Photon
Native vectorized engine for eligible operators — and the UDF fallback that cancels it.
UniForm and Iceberg
One copy of the Parquet, two table formats on top. Delta UniForm and managed Iceberg tables.
Row filters and column masks
Enforce who sees which rows and which columns inside the table, not in every query.
Delta Sharing and Clean Rooms
Share live tables across clouds and platforms without copying them, and compute on data you never see.
System tables
Billing, lineage, audit, and query history as queryable tables — the answer to 'who spent that?'
Predictive Optimization
Stop scheduling OPTIMIZE and VACUUM. Databricks decides which tables actually need maintenance.
Serverless compute
Warm pools, seconds to start, and a bill that stops with the query — plus what you give up.
Streaming tables and materialized views
Declare the result you want in SQL. Incremental refresh keeps it current without a pipeline.
Lakeflow
Connect, Declarative Pipelines, and Jobs as one ingestion-to-orchestration surface.
Asset Bundles and CI/CD
Your jobs, pipelines, and notebooks as YAML in git, deployed per environment from CI.
AI functions and Vector Search
Call a model from SQL with ai_query, and keep an index in sync with Vector Search.
