I85PREP
⌘ K
IDMC for Databricks Foundation study materials
IDMC for Databricks Foundation study resources

IDMC for Databricks Foundation practice questions

One question at a time. Build your IDMC for Databricks Foundation understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1IDMC Platform Fundamentals and Architecture

In Informatica Intelligent Data Management Cloud (IDMC), which layer is responsible for storing metadata about assets such as tables, files, and pipelines so that they can be discovered and governed?

  1. The Metadata Management layer, which captures technical, business, and operational metadata in a central catalog Correct answer.
  2. The Compute layer, which executes mapping logic on the Secure Agent
  3. The Connectivity layer, which brokers JDBC sessions to source systems
  4. The Presentation layer, which renders the user interface in the browser

Explanation

IDMC is organized as a set of cloud services built around a shared metadata backbone. Metadata Management captures technical, business, and operational metadata so assets can be catalogued, searched, and governed across services. Compute and connectivity layers move and process data but do not own the governed metadata repository.

Question 2IDMC Platform Fundamentals and Architecture

A customer wants to use IDMC Cloud Data Integration without installing any software on their own machines. Which statement describes the supported runtime model?

  1. All mappings must run on a Secure Agent installed on a customer-hosted machine.
  2. Mappings can run on the cloud-hosted serverless runtime, with Secure Agents reserved for private-network or on-premises access. Correct answer.
  3. Mappings can only run inside the Databricks cluster through a Spark submission.
  4. Mappings require a dedicated gateway appliance that Informatica ships to the customer site.

Explanation

Cloud Data Integration offers a serverless, Informatica-hosted runtime for cloud sources and targets. A Secure Agent is needed when data or endpoints sit behind a firewall or in a private network, because the agent provides outbound connectivity and local execution. The serverless option removes the need for customer-managed infrastructure in the simple case.

Question 3IDMC Platform Fundamentals and Architecture

Which IDMC capability provides a business-friendly, searchable inventory of data assets with lineage, ratings, and glossary linkage?

  1. Cloud Data Governance and Catalog Correct answer.
  2. Cloud Data Quality
  3. Cloud Application Integration
  4. Cloud B2B Gateway

Explanation

Cloud Data Governance and Catalog is the IDMC service that catalogs assets, captures lineage, and connects technical metadata to business glossaries. It is the discovery and governance surface for the platform. The other services focus on integration, quality processing, or partner messaging rather than cataloging.

Question 4IDMC Platform Fundamentals and Architecture

In the IDMC organizational model, what is the purpose of an organization?

  1. It is a Databricks workspace attached to a metastore.
  2. It is the physical server that hosts the Informatica Cloud services.
  3. It is a single mapping that can be scheduled independently.
  4. It is a logical container that owns users, licenses, and sub-organizations such as environments. Correct answer.

Explanation

An IDMC organization is the top-level tenant that holds users, roles, licenses, and subordinate structures such as environments. Environments then separate development, test, and production content. Mappings and Databricks workspaces are content or external systems, not the organizational container.

Question 5IDMC Platform Fundamentals and Architecture

Which component is installed in a customer environment to let IDMC tasks reach on-premises databases and file systems?

  1. The Metadata Command Center
  2. The Data Integration Hub
  3. The Secure Agent Correct answer.
  4. The Cloud Connector Framework

Explanation

The Secure Agent is lightweight software installed on a customer machine or VM. It establishes outbound connections to IDMC and executes tasks close to private data sources. Catalog and hub services are cloud-side capabilities rather than local connectivity software.

Question 6IDMC Platform Fundamentals and Architecture

A team needs to connect IDMC Cloud Data Integration to a Databricks SQL warehouse as a target. Which connection property is essential for the connection to authenticate and address the correct compute?

  1. The Hive metastore Thrift port and a Kerberos keytab
  2. The Spark master URL and a cluster driver private IP
  3. The Delta Lake transaction log path on cloud object storage
  4. The Databricks workspace URL, an HTTP path for the SQL warehouse, and a personal access token or OAuth credential Correct answer.

Explanation

Databricks connections in IDMC use the workspace server hostname, the HTTP path that identifies the SQL warehouse or cluster, and a token or OAuth credential for authentication. Those three elements let the runtime route queries to the right compute. Metastore ports, Spark master URLs, and raw log paths are not the connection contract.

Question 7IDMC Platform Fundamentals and Architecture

An administrator must ensure that developers in one IDMC environment cannot publish assets directly to production. Which mechanism is designed for this separation?

  1. Using environments with role-based privileges and a promotion process between them Correct answer.
  2. Creating separate organizations for each developer
  3. Relying on Databricks cluster access modes alone
  4. Disabling the Secure Agent in production

Explanation

IDMC environments partition content and are governed by roles and privileges, so developers can be restricted to non-production environments. Assets are then promoted through a controlled migration process. Per-developer organizations and agent removal do not implement a governed promotion path.

Question 8IDMC Platform Fundamentals and Architecture

Which IDMC service would you choose to define reusable data quality rules, run profiling, and score records before loading them into a Databricks table?

  1. Cloud B2B Gateway
  2. Cloud Application Integration
  3. Cloud API Management
  4. Cloud Data Quality Correct answer.

Explanation

Cloud Data Quality provides profiling, rule specification, and scorecards that assess and remediate record-level quality. Those rules can be invoked in integration mappings before data lands in a target such as Databricks. The other services address process orchestration, API exposure, or partner messaging.

Question 9IDMC Platform Fundamentals and Architecture

A data steward wants to see how a column in a Databricks table was derived from upstream source fields. Which IDMC capability delivers this?

  1. Row-level filtering in a mapping
  2. Column-level lineage in Cloud Data Governance and Catalog Correct answer.
  3. Elastic provisioning of the serverless runtime
  4. Taskflow scheduling in Cloud Data Integration

Explanation

Cloud Data Governance and Catalog captures lineage, including column-level relationships, by analyzing metadata from integration jobs and scanned sources. This lets stewards trace derivation across systems. Scheduling, runtime scaling, and row filtering are execution concerns that do not produce lineage views.

Question 10IDMC Platform Fundamentals and Architecture

Which statement accurately describes how IDMC Cloud Data Integration maps to Databricks as a target during execution?

  1. IDMC bypasses Databricks and writes directly to the underlying cloud storage tables.
  2. IDMC can push mapping logic to the Databricks cluster so transformation runs in Spark, using the connection to submit and monitor the job. Correct answer.
  3. IDMC generates Spark code that runs only on the Secure Agent's local JVM.
  4. IDMC writes CSV files to DBFS and requires a separate Databricks job to load them.

Explanation

The Databricks connection supports pushdown, where IDMC generates and submits Spark logic to run on Databricks compute. The runtime then monitors the job and reports status back to IDMC. This uses Databricks capacity rather than processing everything locally on the agent.

Question 11IDMC Platform Fundamentals and Architecture

Which IDMC capability lets an organization expose a data integration process as a REST API for external consumers?

  1. Cloud Data Governance and Catalog glossary
  2. Cloud Data Integration mapping task
  3. Cloud Data Quality scorecard
  4. Cloud Application Integration process exposed as a service Correct answer.

Explanation

Cloud Application Integration builds processes and services that can be published as REST or SOAP endpoints. It can call data integration tasks and orchestrate multiple systems behind a single API. Mapping tasks, scorecards, and glossaries are not API publication mechanisms.

Question 12IDMC Platform Fundamentals and Architecture

What is the primary role of the Metadata Command Center in an IDMC deployment?

  1. It stores the encrypted credentials used by Secure Agents.
  2. It scans and ingests metadata from external sources and catalogs into IDMC. Correct answer.
  3. It compiles mappings into Spark code.
  4. It schedules Secure Agent upgrades.

Explanation

Metadata Command Center performs metadata extraction and ingestion from connected sources so that assets populate the catalog. It works with scanners and connection resources to bring technical metadata into governance. Credential storage, agent upgrades, and code compilation are separate platform functions.

Question 13IDMC Platform Fundamentals and Architecture

Which IDMC licensing concept determines which services and features a user can access within an organization?

  1. The entitlement or service tier assigned to the organization Correct answer.
  2. The mapping parameter file
  3. The Secure Agent group
  4. The Databricks cluster policy

Explanation

Access to IDMC services is governed by the entitlements or service tiers provisioned for the organization. Those entitlements determine which capabilities appear and can be used. Agent groups, cluster policies, and parameter files are operational constructs that do not grant service access.

Question 14IDMC Platform Fundamentals and Architecture

A security architect asks how data moves between a Secure Agent and IDMC cloud services. Which description is correct?

  1. The agent opens outbound HTTPS connections to IDMC, so no inbound firewall rule is required. Correct answer.
  2. The agent communicates only through email notifications.
  3. The agent requires a site-to-site VPN with inbound SSH access.
  4. The cloud service opens inbound connections to the agent on a fixed public port.

Explanation

Secure Agents initiate outbound connections to IDMC over HTTPS, which avoids exposing inbound ports in the customer network. This outbound-only pattern simplifies firewall configuration and reduces the attack surface. Inbound access, VPN tunnels, and email transport are not part of the standard model.

Question 15IDMC Platform Fundamentals and Architecture

Which pairing correctly matches an IDMC service to its core purpose?

  1. Cloud Data Integration - cataloging business glossaries
  2. Cloud Data Quality - orchestrating REST APIs
  3. Cloud Data Integration - executing mappings and tasks that move and transform data Correct answer.
  4. Cloud B2B Gateway - profiling column completeness

Explanation

Cloud Data Integration is the service that runs mappings and taskflows to move and transform data across sources and targets. Glossaries belong to governance, API orchestration to application integration, and profiling to data quality. Matching each service to its purpose is essential for architecture decisions.

Question 16IDMC Platform Fundamentals and Architecture

An organization needs near-real-time ingestion of change events from a Databricks Delta table into IDMC-managed downstream systems. Which architectural consideration matters most?

  1. Relying on catalog scans to move the data
  2. Using a batch mapping scheduled hourly with full reloads
  3. Configuring a streaming or change-data-capture capable source and a continuously running task Correct answer.
  4. Disabling the Secure Agent to reduce latency

Explanation

Near-real-time propagation requires a source mechanism that emits changes, such as change data capture or streaming reads, plus a task that runs continuously rather than on a batch schedule. Batch full reloads introduce latency and load. Catalog scans observe metadata and never move data.

Question 17IDMC Platform Fundamentals and Architecture

Which statement best explains why IDMC separates the metadata layer from the execution runtime?

  1. It guarantees that all mappings execute on the Secure Agent.
  2. It removes the need for authentication in Databricks connections.
  3. It allows governance, lineage, and discovery to operate on assets independently of where or when processing runs. Correct answer.
  4. It forces every organization to use a single environment.

Explanation

Decoupling metadata from execution means cataloging, lineage, and governance persist even as jobs run on serverless, agent-based, or Databricks compute. This supports consistent oversight across heterogeneous runtimes. It does not dictate a single runtime, remove authentication, or collapse environments.

Question 18IDMC Platform Fundamentals and Architecture

A team observes that a mapping pushed to Databricks fails only when run in production, while the same mapping succeeds in development. Which difference is the most plausible architectural cause?

  1. The production connection points to a different workspace or warehouse with different permissions or compute configuration. Correct answer.
  2. The catalog contains more assets in production.
  3. The development environment uses a different glossary term.
  4. The organization has more users in production.

Explanation

Pushdown execution depends on the connection's target workspace, warehouse, and the identity's privileges on that compute and its catalog objects. A production connection with different endpoints or grants can fail where development succeeds. Glossary terms, catalog size, and user counts do not determine mapping execution success.

Question 19IDMC Platform Fundamentals and Architecture

Which design decision most directly reduces data egress and latency when integrating large Databricks tables with IDMC?

  1. Pushing transformation logic to Databricks so processing occurs where the data resides Correct answer.
  2. Storing intermediate results in the glossary
  3. Pulling all rows to the Secure Agent and transforming them locally
  4. Increasing the number of catalog scans per hour

Explanation

Pushdown keeps large-scale processing on Databricks compute next to the data, avoiding bulk transfer to an agent and back. This lowers egress cost and round-trip latency. Local transformation of full tables and metadata-only activities do not achieve these savings.

Question 20IDMC Platform Fundamentals and Architecture

An architect must decide between serverless runtime and a Secure Agent for a cloud-to-cloud pipeline that reads from Databricks and writes to a cloud object store. Which reasoning is sound?

  1. Serverless cannot connect to cloud object stores.
  2. The choice has no impact because both runtimes execute on Databricks.
  3. Serverless is preferable because no private-network access is needed, reducing operational overhead. Correct answer.
  4. A Secure Agent is mandatory because Databricks sources always require local execution.

Explanation

When both endpoints are reachable over the internet and no private network is involved, the serverless runtime avoids managing agent infrastructure. A Secure Agent is warranted when private connectivity or local execution is required. Serverless does support cloud storage targets, and the runtimes are distinct.

Question 21IDMC Platform Fundamentals and Architecture

Which IDMC concept groups Secure Agents so that tasks can be routed to a specific set of machines?

  1. Environment
  2. Glossary
  3. Agent group Correct answer.
  4. Connection

Explanation

An agent group is a logical collection of Secure Agents that tasks can target, enabling routing and load distribution. Environments partition content, glossaries define terms, and connections hold endpoint details. Grouping agents is how administrators control which machines execute work.

Question 22IDMC Platform Fundamentals and Architecture

Which statement about IDMC roles is accurate?

  1. Roles bundle privileges that determine what a user can view or change within the organization. Correct answer.
  2. Roles are assigned to Databricks clusters to control Spark executors.
  3. Roles define the physical location of the Secure Agent.
  4. Roles are synonyms for environments.

Explanation

IDMC roles are collections of privileges granted to users so administrators can control access to services and assets. They are distinct from environments, which partition content, and from Databricks cluster settings. Agent location is a deployment detail unrelated to role definition.

Question 23IDMC Platform Fundamentals and Architecture

Which IDMC feature supports defining business terms and linking them to technical columns so analysts understand data meaning?

  1. Business glossary in Cloud Data Governance and Catalog Correct answer.
  2. Mapping parameter in Cloud Data Integration
  3. Scorecard in Cloud Data Quality
  4. Taskflow in Cloud Data Integration

Explanation

The business glossary in Cloud Data Governance and Catalog captures definitions and associates them with technical assets such as columns. This bridges business vocabulary and physical schemas. Taskflows, parameters, and scorecards serve execution or measurement rather than term definition.

Question 24IDMC Platform Fundamentals and Architecture

What distinguishes Cloud Data Integration taskflows from individual mapping tasks in IDMC?

  1. Taskflows orchestrate multiple tasks and services in a defined sequence with dependencies. Correct answer.
  2. Taskflows are only used to store glossary terms.
  3. Taskflows replace the Secure Agent.
  4. Taskflows cannot call data quality rules.

Explanation

A taskflow coordinates several tasks and service steps, including dependencies and conditional flow, into an end-to-end process. Individual mapping tasks execute a single mapping. Taskflows do not store glossary content, replace agents, or exclude quality rules.

Question 25IDMC Platform Fundamentals and Architecture

A governance lead wants automated detection of sensitive data such as personal identifiers in Databricks tables before granting broad analyst access. Which IDMC combination is most appropriate?

  1. Relying on Databricks cluster autoscaling
  2. Catalog scanning with data classification and quality profiling to identify and tag sensitive columns Correct answer.
  3. Using B2B Gateway to mask EDI payloads
  4. Disabling lineage to reduce metadata volume

Explanation

Catalog scanning ingests metadata, classification identifies sensitive patterns, and profiling validates content so columns can be tagged and governed. Together they support access decisions. Disabling lineage removes useful context, and the other options do not detect sensitive data in tables.

Question 26IDMC Platform Fundamentals and Architecture

An architect is explaining the IDMC architecture to a Databricks engineering team. Which statement accurately describes the division of responsibility between the Informatica cloud services and the Secure Agent runtime?

  1. The cloud services host design-time metadata, orchestration, and monitoring, while the Secure Agent executes jobs and handles connectivity to data sources and targets. Correct answer.
  2. The Secure Agent and cloud services both execute mappings, and the repository is replicated to Databricks so that Databricks schedules all jobs.
  3. The cloud services execute all mappings in a shared multi-tenant cluster, while the Secure Agent only stores repository backups and audit logs.
  4. The Secure Agent hosts the repository of mappings and taskflows, while the cloud services only provide single sign-on and user provisioning.

Explanation

IDMC splits control plane from data plane: the cloud-hosted services hold design metadata, orchestration, and monitoring, while the Secure Agent runs on customer infrastructure and performs the actual data movement and connectivity. This separation lets the agent reach private networks and on-premises systems that cloud services cannot access directly.

Question 27IDMC Platform Fundamentals and Architecture

A data engineer is choosing between an Elastic Serverless runtime and a Secure Agent for a pipeline that must read from an on-premises Oracle database and write to a Databricks SQL warehouse. Which consideration should drive the decision?

  1. Serverless runtime can reach any network endpoint, so it is always preferred for hybrid pipelines.
  2. Serverless runtime and Secure Agent differ only in cost, with identical network reachability.
  3. The Secure Agent cannot write to a Databricks SQL warehouse, so pushdown must be used instead.
  4. Only the Secure Agent can connect to databases inside a private corporate network, so the on-premises Oracle source requires an agent. Correct answer.

Explanation

An Elastic Serverless runtime runs in Informatica-managed cloud infrastructure and therefore cannot open connections into a private corporate network. Because the Oracle source is on-premises, a Secure Agent installed inside that network is required to reach it, after which the mapping can still target Databricks.

Question 28IDMC Platform Fundamentals and Architecture

An administrator has created a new user in IDMC and needs the user to build mappings in Cloud Data Integration but must prevent that user from administering the organization. Which mechanism enforces this?

  1. Assigning the user to a runtime environment, which controls which services appear in the navigation menu.
  2. Assigning the user a personal connection, which restricts them to read-only access across all services.
  3. Assigning the user to a Secure Agent group, which automatically limits the user to design-only operations.
  4. Assigning the user to a role that grants the Designer privilege for Cloud Data Integration while excluding administrative privileges. Correct answer.

Explanation

Access in IDMC is governed by roles, which bundle service-specific privileges such as Designer or Admin. Granting the Designer privilege for Cloud Data Integration without administrative privileges lets the user build mappings while blocking organization-level administration.

Question 29IDMC Platform Fundamentals and Architecture

A governance team wants to scan Databricks tables, catalog technical metadata, and link those columns to business glossary terms so that analysts can trace definitions. Which combination of IDMC capabilities fulfills this requirement?

  1. Cloud Application Integration processes that call the Databricks REST API to write term definitions into table comments.
  2. Cloud Data Integration mappings plus taskflows that copy glossary terms into Databricks tables.
  3. Metadata Command Center for metadata extraction and cataloging, together with the business glossary for term-to-column linkage. Correct answer.
  4. Data Quality rule specifications alone, because quality scores imply the business meaning of each column.

Explanation

Metadata Command Center extracts and catalogs technical metadata from sources such as Databricks and supports linking catalog assets to glossary terms. The business glossary supplies the shared definitions, enabling lineage and definition traceability for analysts.

Question 30IDMC Platform Fundamentals and Architecture

An architect is designing an IDMC pipeline that reads a large Delta table in Databricks, applies a filter and aggregation, and writes results back to another Databricks table. To minimize data movement out of the lakehouse, which design approach is most appropriate?

  1. Split the table into many small mappings so that each mapping runs concurrently on separate agents.
  2. Convert the pipeline to a Cloud Application Integration process that streams rows through the IDMC cloud services.
  3. Use a mapping in native Databricks mode so the transformation logic is pushed down and executed by the Databricks engine rather than the agent. Correct answer.
  4. Extract the full Delta table to the Secure Agent, transform it there, and load the results back through a staging file.

Explanation

Native Databricks execution pushes transformation logic into the Databricks engine, so the large table is filtered and aggregated where it resides instead of being moved to the agent. This reduces egress, network latency, and agent resource consumption.

Question 31Databricks Lakehouse Concepts and Integration Patterns

In the Databricks Lakehouse architecture, which layer is responsible for storing raw data exactly as it was ingested and typically serves as the source of truth for reprocessing?

  1. The gold layer
  2. The bronze layer Correct answer.
  3. The feature store
  4. The serving layer

Explanation

The medallion architecture organizes data into bronze, silver, and gold layers. The bronze layer holds raw ingested data in its original form and acts as the durable source of truth for downstream reprocessing.

Question 32Databricks Lakehouse Concepts and Integration Patterns

A data engineer wants ACID transactions and time travel on a table stored in cloud object storage. Which Databricks component provides these capabilities?

  1. A Hive metastore
  2. Parquet files
  3. Delta Lake Correct answer.
  4. DBFS mounts

Explanation

Delta Lake adds a transaction log on top of Parquet files, enabling ACID transactions, schema enforcement, and time travel. Plain Parquet files and metastore or mount configurations do not provide transactional guarantees on their own.

Question 33Databricks Lakehouse Concepts and Integration Patterns

Which statement best describes the role of the Delta transaction log (_delta_log) in a Delta table?

  1. It records committed actions such as adding or removing files to define table state Correct answer.
  2. It caches query results so repeated SQL runs skip computation
  3. It holds user credentials used to access the underlying cloud storage
  4. It stores a compressed copy of every row for fast full-table scans

Explanation

The _delta_log directory contains JSON commit files that record actions like add and remove, plus checkpoints. Readers reconstruct the current table snapshot from this log, which is what enables ACID guarantees and time travel.

Question 34Databricks Lakehouse Concepts and Integration Patterns

A team ingests streaming events with Auto Loader into a bronze Delta table. Which Auto Loader capability ensures that each input file is processed only once across runs?

  1. Checkpointing of discovered file state Correct answer.
  2. File notification mode
  3. Trigger.AvailableNow execution
  4. Schema inference with a sampled schema hint

Explanation

Auto Loader tracks which files have been discovered and processed in a checkpoint location, so restarts do not reprocess the same files. Schema inference and trigger settings affect structure and scheduling, not exactly-once file tracking.

Question 35Databricks Lakehouse Concepts and Integration Patterns

Which Delta Lake feature lets an analyst query a table as it existed at a specific version number?

  1. Optimize with file compaction
  2. Z-ordering on a timestamp column
  3. Change data feed
  4. Time travel using VERSION AS OF Correct answer.

Explanation

Delta time travel queries historical snapshots by version or timestamp, for example SELECT * FROM t VERSION AS OF 5. The other features improve layout or expose row-level changes but do not return a prior table snapshot.

Question 36Databricks Lakehouse Concepts and Integration Patterns

A pipeline must merge late-arriving updates into a silver Delta table keyed by order_id. Which operation is the most appropriate?

  1. MERGE INTO the target using the staging updates on order_id Correct answer.
  2. INSERT OVERWRITE the entire target table from staging
  3. CREATE OR REPLACE TABLE from the staging data
  4. TRUNCATE the target and reload all rows

Explanation

MERGE INTO performs upserts by matching on a key and applying matched updates and unmatched inserts. Overwriting or truncating the table discards existing rows and is not a targeted upsert pattern.

Question 37Databricks Lakehouse Concepts and Integration Patterns

Which statement about Delta Lake schema enforcement is accurate?

  1. Schema enforcement requires the table to be partitioned
  2. Extra columns are always silently added to the target schema
  3. Schema enforcement applies only to streaming writes
  4. Writes with extra columns are rejected unless schema evolution is explicitly enabled Correct answer.

Explanation

By default Delta rejects writes whose schema does not match the table, protecting data quality. Schema evolution must be opted into, for example with mergeSchema, before new columns are added automatically.

Question 38Databricks Lakehouse Concepts and Integration Patterns

A data engineer needs a single governance layer that controls access to tables, files, and notebooks across workspaces. Which Databricks capability addresses this?

  1. Delta Live Tables
  2. Unity Catalog Correct answer.
  3. DBFS root storage
  4. Cluster-scoped init scripts

Explanation

Unity Catalog provides centralized governance for data and AI assets, including fine-grained privileges across workspaces. DBFS roots, init scripts, and pipeline frameworks do not deliver cross-workspace access control.

Question 39Databricks Lakehouse Concepts and Integration Patterns

Which table type in Unity Catalog is best suited for high-volume fact tables that are frequently filtered on a timestamp column?

  1. A managed table partitioned by the timestamp column Correct answer.
  2. An external table with no partitioning
  3. A view over a raw bronze table
  4. A temporary view scoped to a session

Explanation

Partitioning a large fact table on a common filter column lets queries prune irrelevant files. Unpartitioned external tables, views, and temporary views do not reduce the scanned data volume for these filters.

Question 40Databricks Lakehouse Concepts and Integration Patterns

What is the primary purpose of the OPTIMIZE command with Z-ORDER on a Delta table?

  1. To convert the table into a Parquet-only format
  2. To compact small files and co-locate related data for better data skipping Correct answer.
  3. To delete rows that violate a check constraint
  4. To enforce column constraints on future writes

Explanation

OPTIMIZE rewrites many small files into fewer larger ones, and Z-ORDER clusters rows by chosen columns so file statistics enable skipping. Constraints and format conversion are handled by other commands.

Question 41Databricks Lakehouse Concepts and Integration Patterns

A streaming query reads from a Delta table and must keep processing new rows as they are appended. Which source configuration is required?

  1. Export the table to CSV and stream from the export
  2. Read the table with read and cache the result
  3. Query the table through a temporary view
  4. Read the table with readStream and treat it as a streaming source Correct answer.

Explanation

Delta tables can act as streaming sources when read with readStream, allowing incremental consumption of new commits. Batch reads, views, and exports do not provide continuous incremental processing.

Question 42Databricks Lakehouse Concepts and Integration Patterns

Which statement correctly distinguishes a managed table from an external table in Unity Catalog?

  1. Dropping a managed table deletes its data files, while dropping an external table leaves them in place Correct answer.
  2. Managed tables cannot use Delta Lake as their format
  3. Managed tables always store data in DBFS root
  4. External tables cannot be queried with SQL

Explanation

For managed tables, the catalog owns the data lifecycle, so dropping the table removes the underlying files. External tables reference a location the user controls, so dropping the table removes only metadata.

Question 43Databricks Lakehouse Concepts and Integration Patterns

An engineer wants to capture row-level inserts, updates, and deletes from a Delta table for downstream consumers. Which feature should be enabled?

  1. Change data feed Correct answer.
  2. Time travel retention
  3. Deletion vectors
  4. Photon acceleration

Explanation

Change data feed records row-level change events that downstream queries can read for incremental processing. Time travel retention, deletion vectors, and Photon affect history, storage, and query speed rather than change capture.

Question 44Databricks Lakehouse Concepts and Integration Patterns

A data engineer is designing a gold-layer table for BI dashboards that must reflect near-real-time updates from a silver streaming table. Which pattern best fits the lakehouse?

  1. Create a temporary view over the silver table for each dashboard
  2. Write the silver stream into a gold Delta table using a streaming write with checkpointing Correct answer.
  3. Point the BI tool directly at the bronze ingestion files
  4. Rebuild the gold table nightly with a full overwrite

Explanation

Streaming the silver table into gold with a checkpointed write propagates updates incrementally and preserves exactly-once processing. Full overwrites, raw bronze access, and per-dashboard temporary views do not deliver reliable near-real-time curated data.

Question 45Databricks Lakehouse Concepts and Integration Patterns

Which Delta Lake capability allows a table to accept a changed column type, such as widening an integer column to bigint, during a write?

  1. Z-ordering by column name
  2. Type widening controlled by a table property Correct answer.
  3. Schema enforcement with no options
  4. Liquid clustering

Explanation

Delta supports controlled type widening when the table property is enabled, letting compatible type changes be applied to existing data. Plain schema enforcement rejects mismatches, and clustering features affect layout rather than types.

Question 46Databricks Lakehouse Concepts and Integration Patterns

What does the Delta Lake VACUUM command do?

  1. It compacts small files into larger ones for faster reads
  2. It recomputes statistics used by the query optimizer
  3. It restores a table to an earlier version
  4. It removes data files no longer referenced by the transaction log beyond the retention threshold Correct answer.

Explanation

VACUUM deletes unreferenced files older than the configured retention period, reclaiming storage. Compaction is handled by OPTIMIZE, statistics by ANALYZE, and restoration by RESTORE.

Question 47Databricks Lakehouse Concepts and Integration Patterns

A team needs to enforce that a column in a Delta table is never null on future writes. Which mechanism should they use?

  1. A change data feed setting
  2. A NOT NULL constraint added with ALTER TABLE Correct answer.
  3. A Z-ORDER on that column
  4. A checkpoint on the streaming query

Explanation

Delta supports constraints such as NOT NULL and CHECK, added via ALTER TABLE, which reject violating writes. Clustering, change data feed, and checkpoints do not enforce column-level data quality.

Question 48Databricks Lakehouse Concepts and Integration Patterns

Which statement best explains why the lakehouse architecture can serve both BI and machine-learning workloads on one copy of data?

  1. Proprietary formats force each workload to maintain a separate copy
  2. ML frameworks cannot read data stored in cloud object storage
  3. BI tools require data to be exported to a relational warehouse first
  4. Open formats plus a transactional layer support SQL engines and ML frameworks over shared storage Correct answer.

Explanation

The lakehouse stores data in open formats with a transactional layer, so SQL engines and ML libraries can read the same governed data without duplication. Separate copies, mandatory exports, and storage limitations contradict this design.

Question 49Databricks Lakehouse Concepts and Integration Patterns

What is the role of a checkpoint location in a Structured Streaming query writing to Delta Lake?

  1. It holds the schema of the source for validation only
  2. It defines the partitioning scheme of the target table
  3. It stores progress and state so the query can recover exactly-once across restarts Correct answer.
  4. It stores a copy of the target table for disaster recovery

Explanation

The checkpoint records offsets, commit metadata, and state so a restarted stream resumes without duplicating or dropping data. It is not a table backup, partitioning definition, or schema-only artifact.

Question 50Databricks Lakehouse Concepts and Integration Patterns

An engineer must combine a large Delta fact table with a small dimension table in a streaming pipeline. Which approach best avoids shuffling the large side repeatedly?

  1. Convert both sides to CSV and join externally
  2. Use a stream-static join with the small dimension broadcast Correct answer.
  3. Repartition the fact table on every microbatch
  4. Use a full outer join without hints

Explanation

A stream-static join against a small dimension lets Spark broadcast the static side, avoiding repeated shuffles of the large stream. Unhinted full joins, external CSV joins, and forced repartitioning add cost without this benefit.