Data Intelligence Platform
11%of exam
Unity CatalogCatalog ExplorerMarketplaceLineage Graph
Managing Data
8%of exam
Certified TagsData CleaningLineage CheckCOALESCE/NULLIF
Importing Data
5%of exam
Auto LoaderCOPY INTODelta SharingMarketplace Access
Executing Queries
20%of exam
SQL WarehousesTables vs ViewsJoins + PIVOTTime Travel
Analyzing Queries
15%of exam
Query HistoryQuery ProfileCachingLiquid Clustering
Dashboards + Visualizations
16%of exam
ParametersScheduled RefreshAlertsEmbedding
AI/BI Genie Spaces
12%of exam
InstructionsTrusted AssetsSample QuestionsBenchmarks
Data Modeling
5%of exam
Star SchemaSnowflake SchemaData VaultMedallion
Securing Data
8%of exam
RBACRow FiltersColumn MasksNamespace
Quick Facts
- Exam
- Databricks DA
- Credential
- Data Analyst Associate
- Questions
- 45 scored
- Time
- 90 min
- Pass
- 70%
- Fee
- $200
- Validity
- 2 years
- Blueprint
- Oct 30 2025
Unity Catalog Namespace Order
Always catalog, then schema, then table
Catalog: top-level containerSchema: groups tablesTable: the actual object
Unity Catalog Hierarchy
- Metastore
- Top-level UC container
- Catalog
- Groups schemas together
- Schema
- Groups tables and views
- Table
- Structured data object
- View
- Saved query logic
- Volume
- Non-tabular file storage
- Function
- Registered SQL or Python logic
- Model
- Registered ML model
Discovery + Governance Tools
- Catalog Explorer
- Browse schemas, tables, permissions
- Marketplace
- Discover shared data listings
- Lineage graph
- Trace upstream/downstream dependencies
- Data Intelligence Platform
- Unified lakehouse plus AI
- System tables
- Built-in operational metadata tables
- Search
- Find governed data assets
Managing Certified Data
- Certified tag
- Steward-reviewed trusted dataset
- Custom tag
- Searchable metadata label
- COALESCE
- Replace null with value
- NULLIF
- Return null if equal
- REGEXP_REPLACE
- Pattern-based string cleanup
- Lineage check
- Verify upstream source table
Delta Sharing vs Marketplace
Delta Sharing
- Direct point-to-point share
- Works cross-platform
Marketplace
- Public or private listings
- Browse and subscribe
Private link vs public catalog
Data Ingestion Methods
- UI upload
- Manual file import
- Auto Loader
- Incremental cloud-file ingestion
- COPY INTO
- SQL-based file loading
- Delta Sharing
- Cross-platform secure data share
- API intake
- Programmatic data ingestion
- Marketplace access
- Subscribe to shared listings
Warehouse Startup Speed Order
Serverless seconds, Pro minutes, Classic slowest
Serverless: starts in secondsPro: about four minutesClassic: entry-level only
Materialized View vs Streaming Table
Materialized View
- Batch semantics
- Recomputes or refreshes incrementally
- Best for Silver/Gold layer
Streaming Table
- Streaming semantics
- Row seen exactly once
- Best for Bronze ingest
Correctness vs continuous ingest
Which View Type to Use
- Need governed base table→Managed table
- Data lives outside UC→External table
- Simple saved query logic→View
- Expensive BI-ready aggregation→Materialized view(Batch refresh)
- Continuous append-only ingest→Streaming table
- Row or column security→Dynamic view
SQL Warehouse Types
- Serverless
- Starts in seconds
- Pro
- ~4-minute startup, no IWM
- Classic
- Entry-level performance only
- Photon
- Vectorized engine, all tiers
- Predictive I/O
- Smart file and row skipping
- IWM
- AI-managed serverless scaling
- Auto-stop
- Idle warehouse shuts down
Managed vs External Table
Managed Table
- UC owns the files
- DROP deletes underlying data
- Simplifies governance
External Table
- You own storage path
- DROP keeps underlying files
- Needs external location
UC-owned vs self-owned
Which SQL Warehouse to Pick
- Bursty unpredictable analyst load→Serverless warehouse
- Steady workload, no IWM need→Pro warehouse
- Basic entry-level testing only→Classic warehouse
- Need fastest possible startup→Serverless warehouse(Seconds not minutes)
Tables and Views
- Managed table
- UC owns underlying files
- External table
- You own storage location
- View
- Recomputed every query
- Materialized view
- Precomputed, batch-refreshed result
- Streaming table
- Append-only incremental ingest
- Dynamic view
- Row/column security logic
Serverless vs Pro Warehouse
Serverless
- Starts in seconds
- AI-based IWM scaling
- Lower cost, bursty loads
Pro
- ~4-minute startup
- No IWM scaling
- Better for steady load
Instant vs provisioned compute
SQL Query Techniques
- INNER JOIN
- Matching rows only
- LEFT JOIN
- All left, matched right
- UNION ALL
- Combine rows, keep duplicates
- PIVOT
- Rotate values into columns
- VERSION AS OF
- Delta time travel by version
- FILTER
- Higher-order array filter
- TRANSFORM
- Higher-order array mapper
Query History vs Query Profile
Query History
- Lists past query runs
- Shows duration and status
Query Profile
- Shows one execution plan
- Stage-level timing detail
List view vs deep-dive
Diagnosing a Slow Query
- Query already ran once→Check Query History
- Need stage-level timings→Open Query Profile
- Repeated identical query text→Check result cache
- Filter on one huge column→Add Liquid Clustering
- Need Delta write history→Run DESCRIBE HISTORY
- Row count larger than expected→Recheck join keys
Query Performance + Tuning
- Query History
- List of past runs
- Query Profile
- Stage-level execution timings
- Result cache
- Reuses identical prior query
- Disk cache
- Caches remote file reads
- Liquid Clustering
- Flexible, incremental data layout
- DESCRIBE HISTORY
- Delta table audit log
- Spill to disk
- Memory overflow, indicates slowness
Notebook Chart vs AI/BI Dashboard
Notebook Chart
- One-off exploration
- Lives inside one cell
AI/BI Dashboard
- Recurring stakeholder view
- Shareable, scheduled, alerting
Explore vs publish
AI/BI Dashboard Features
- Parameter
- Reusable filter-like input
- Scheduled refresh
- Automatic periodic data update
- Alert
- Notify on threshold breach
- Shareable link
- Controlled external audience access
- Embedding
- Dashboard inside external app
- Multi-page layout
- Organize related visuals together
Genie Quality Improvement Loop
Curate, instruct, sample, benchmark, improve
Curate trusted assets firstAdd clear instructionsProvide sample questionsTrack benchmarks and feedback
Improving a Genie Space
- Business jargon misunderstood→Add instructions
- Users unsure what to ask→Add sample questions
- Answers cite wrong tables→Curate trusted assets
- Need ongoing quality tracking→Review benchmarks, feedback
Genie Space Setup
- Instructions
- Guide Genie's business terms
- Sample questions
- Show users example prompts
- Trusted assets
- Verified queries Genie prioritizes
- Benchmarks
- Track answer quality over time
- Feedback
- Users flag good or bad
- Curated datasets
- Scoped tables Genie can query
Medallion Architecture Layers
Bronze raw, Silver clean, Gold curated
Bronze: raw ingestSilver: cleaned and joinedGold: BI-ready curated
Star vs Snowflake Schema
Star Schema
- Denormalized dimensions
- Fewer joins needed
- Faster BI reads
Snowflake Schema
- Normalized dimensions
- More joins needed
- Less data redundancy
Simple vs normalized
Choosing a Data Model Pattern
- Simple BI star reporting→Star schema
- Normalize shared dimension data→Snowflake schema
- Need full historical auditability→Data vault
- Raw to trusted layering→Medallion architecture
Data Modeling Patterns
- Star schema
- Denormalized facts and dimensions
- Snowflake schema
- Normalized, more joins
- Data vault
- Hubs, links, satellites model
- Fact table
- Measures and foreign keys
- Dimension table
- Descriptive business attributes
RBAC vs Row/Column Masking
RBAC (GRANT)
- Object-level access control
- All-or-nothing per object
Row Filter / Column Mask
- Row-level visibility rules
- Column value obscuring
Access vs visible content
Unity Catalog Security
- Three-level namespace
- Catalog.schema.table addressing
- RBAC
- GRANT/REVOKE object privileges
- Row filter
- Restrict visible rows
- Column mask
- Obscure sensitive column values
- Ownership
- Controls who can grant
- PII protection
- Mask or restrict sensitive data
Common Traps
Managed Table vs External Drop
Managed drop deletes files ≠ External drop keeps files
View vs Materialized View
View recomputes every run ≠ MV stores precomputed result
Cached Result vs Fresh Profile
Cached run skips profiling ≠ Profile needs fresh execution
Certified Tag vs Data Quality
Certified means steward-reviewed ≠ Not a quality guarantee
Dashboard vs Genie Space
Dashboard shows fixed view ≠ Genie answers ad hoc
Photon vs Serverless Warehouse
Photon is query engine ≠ Serverless is compute type
Tag vs Comment Metadata
Tags are searchable metadata ≠ Comments are free text
Last Minute
- 1.45 scored questions in 90 minutes
- 2.Pass at 70% or higher
- 3.SQL execution weighted heaviest at 20%
- 4.Dashboards, queries, Genie exceed half weight
- 5.Managed table drop deletes data
- 6.External table drop keeps files
- 7.Materialized views are batch, precomputed
- 8.Streaming tables are append-only ingest
- 9.Serverless warehouses start in seconds
- 10.Genie needs curated trusted assets
- 11.Row filters and column masks differ
- 12.Namespace order is catalog.schema.table
Buy on AmazonSame family resources
Explore More Databricks Certifications
Continue into nearby exams from the same family. Each card keeps practice questions, study guides, flashcards, videos, and articles in one place.
More From This Family
Videos and articles for deeper review.
VideoDatabricks Certified Data Analyst Associate Exam Guide 2026: SQL, AI/BI Dashboards, Genie, and Free PracticeThis examVideoDatabricks Data Engineer Associate Exam Guide (2026): Pass on Your First TryArticleDatabricks Certified Data Analyst Associate Exam Guide 2026: SQL, AI/BI Dashboards, Genie, and Free Practice14 min readArticleDatabricks Data Engineer Associate Exam Guide (2026): Pass on Your First Try26 min read