I85PREP
⌘ K
IBM C1000-120 study materials
IBM C1000-120 study resources

IBM C1000-120 practice questions

One question at a time. Build your IBM C1000-120 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 1Exam objectives (not enumerated in retrieved source)

An administrator is planning a deployment of IBM Cloud Pak for Data. Which statement best describes the platform's foundational purpose?

  1. It is a standalone desktop reporting tool installed on each analyst workstation
  2. It provides a unified data and AI platform for collecting, organizing, and analyzing data across hybrid cloud environments Correct answer.
  3. It is a network monitoring suite focused exclusively on packet capture
  4. It is a dedicated hardware appliance that replaces on-premises storage arrays

Explanation

Cloud Pak for Data is positioned as an integrated data and AI platform that unifies governance, cataloging, transformation, and model development across hybrid and multicloud estates. Its value comes from consolidating these capabilities rather than replacing storage hardware or acting as a desktop tool.

Question 2Exam objectives (not enumerated in retrieved source)

A team wants to catalog data assets, assign business terms, and track data lineage across sources. Which Cloud Pak for Data capability area addresses this?

  1. Watson Studio notebooks
  2. The data governance and catalog services Correct answer.
  3. The SPSS Modeler visual flow editor
  4. The Cognos dashboard authoring module

Explanation

Cataloging assets, attaching business glossary terms, and viewing lineage are governance functions. These services let stewards publish assets and trace how data moves between systems.

Question 3Exam objectives (not enumerated in retrieved source)

Which component of Cloud Pak for Data acts as the central point for managing users, projects, and service provisioning?

  1. The Db2 warehouse engine
  2. The Watson Machine Learning runtime
  3. The platform control plane and web client Correct answer.
  4. The Data Virtualization query optimizer

Explanation

The control plane hosts the web client where administrators provision services, create projects, and manage access. Other components are services that run under that control plane.

Question 4Exam objectives (not enumerated in retrieved source)

An organization requires that data scientists work with data without copying it into a new repository. Which Cloud Pak for Data service is designed for this?

  1. Watson Knowledge Catalog
  2. Data Virtualization Correct answer.
  3. Cognos Analytics
  4. Watson OpenScale

Explanation

Data Virtualization presents remote sources as virtual tables so queries run against the original systems without physical copies. This reduces duplication and keeps data close to its system of record.

Question 5Exam objectives (not enumerated in retrieved source)

A data steward needs to define a common vocabulary so that 'customer' means the same thing across departments. Which artifact should be created?

  1. A notebook template
  2. A deployment space
  3. A connection asset
  4. A business term in the glossary Correct answer.

Explanation

Business terms in the glossary capture agreed definitions and can be linked to assets so consumers see consistent meaning. Glossary terms are the governance mechanism for shared vocabulary.

Question 6Exam objectives (not enumerated in retrieved source)

Which Cloud Pak for Data service is used to build, train, and deploy machine learning models with tracked experiments?

  1. Db2 Warehouse
  2. Data Virtualization
  3. Watson Knowledge Catalog
  4. Watson Studio and Watson Machine Learning Correct answer.

Explanation

Watson Studio provides the authoring environment for model development, and Watson Machine Learning handles training runs and deployment. Together they cover the model lifecycle from experiment to serving.

Question 7Exam objectives (not enumerated in retrieved source)

A model in production is showing degraded accuracy over time. Which capability helps detect this drift and explain model behavior?

  1. Watson OpenScale monitoring Correct answer.
  2. The project's Git integration
  3. The platform's user management console
  4. The catalog's lineage view

Explanation

OpenScale evaluates deployed models against payload and feedback data to surface accuracy drift, fairness, and explainability insights. It is the monitoring layer for models after they leave the notebook.

Question 8Exam objectives (not enumerated in retrieved source)

Which statement accurately describes how Cloud Pak for Data supports hybrid deployment?

  1. It runs only on a single public cloud region chosen at install time
  2. It is limited to air-gapped environments with no network access
  3. It can be deployed on IBM Cloud, on other clouds, and on customer-managed OpenShift clusters Correct answer.
  4. It requires a dedicated mainframe partition for every project

Explanation

The platform is built to run on Red Hat OpenShift, which allows installation on IBM Cloud, third-party clouds, and on-premises clusters. This portability is central to its hybrid data strategy.

Question 9Exam objectives (not enumerated in retrieved source)

A user needs to query a table in a remote database from within a project without importing the rows. Which asset type should the user create?

  1. A connected data asset referencing a virtualized table Correct answer.
  2. A dashboard visualization
  3. A CSV file uploaded to the project
  4. A notebook with hard-coded sample rows

Explanation

A connected data asset points to a virtualized table so queries are pushed to the source system rather than copying rows into the project. This preserves the source as the system of record.

Question 10Exam objectives (not enumerated in retrieved source)

Which practice best supports reproducibility of a data science project in Cloud Pak for Data?

  1. Disabling asset versioning to reduce storage use
  2. Associating the project with a Git repository and pinning environment versions Correct answer.
  3. Emailing notebooks to teammates as attachments
  4. Storing intermediate results only in local notebook memory

Explanation

Linking a project to Git tracks changes to notebooks and code, while pinned environments ensure the same libraries are available on rerun. Together they let another analyst recreate results reliably.

Question 11Exam objectives (not enumerated in retrieved source)

An administrator is configuring access so that analysts can view cataloged assets but cannot modify glossary terms. Which mechanism enforces this?

  1. The choice of storage class for the Db2 instance
  2. Role-based access control on platform roles and collaborator permissions Correct answer.
  3. The number of CPUs allocated to the analytics engine
  4. The retention period configured on the audit log

Explanation

Platform roles and per-asset collaborator settings determine who may view, edit, or administer resources. Assigning a viewer role limits analysts to read access while stewards retain editing rights.

Question 12Exam objectives (not enumerated in retrieved source)

Which Cloud Pak for Data capability helps an organization discover which downstream reports depend on a specific source column?

  1. Connection pooling
  2. Notebook scheduling
  3. Lineage analysis Correct answer.
  4. Model drift scoring

Explanation

Lineage captures relationships between assets as data flows through transformations and reports. Tracing lineage backward or forward reveals the downstream consumers of a given column.

Question 13Exam objectives (not enumerated in retrieved source)

What is the primary role of a deployment space in Cloud Pak for Data?

  1. To host the OpenShift cluster's networking configuration
  2. To group deployed models and related assets for runtime serving and monitoring Correct answer.
  3. To store raw source files before cataloging
  4. To define the glossary hierarchy for the organization

Explanation

A deployment space collects deployable assets such as models and functions so they can be served and monitored together. It is the runtime boundary rather than an authoring or governance container.

Question 14Exam objectives (not enumerated in retrieved source)

A team wants to enrich a dataset by matching addresses against a reference table of postal codes. Which approach fits the platform's data transformation tooling?

  1. Manually editing each address in a spreadsheet
  2. Disabling the catalog entry for the reference table
  3. Building a transformation flow that joins the dataset to the reference table Correct answer.
  4. Deleting rows without postal codes to simplify the dataset

Explanation

Transformation flows express joins and enrichment steps visually or through generated code, producing a repeatable pipeline. Joining to the reference table adds the postal attributes while keeping the process auditable.

Question 15Exam objectives (not enumerated in retrieved source)

Which statement about the relationship between projects and catalogs in Cloud Pak for Data is accurate?

  1. Projects and catalogs are the same object with different labels
  2. Assets can be published from a project to a catalog for broader discovery Correct answer.
  3. Catalogs can only contain notebooks, never data assets
  4. Publishing to a catalog deletes the asset from the project

Explanation

Projects are working environments, while catalogs are shared discovery surfaces. Publishing promotes an asset to the catalog so others can find and request access without joining the project.

Question 16Exam objectives (not enumerated in retrieved source)

Which capability allows analysts to run SQL against federated sources without writing connection code for each system?

  1. The Data Virtualization query service Correct answer.
  2. The OpenScale fairness checker
  3. The platform's backup utility
  4. The project's activity log

Explanation

Data Virtualization exposes remote sources as tables that can be queried with standard SQL, handling connection and pushdown behind the scenes. Analysts then write SQL rather than bespoke connectors.

Question 17Exam objectives (not enumerated in retrieved source)

A model must be deployed so that business applications can call it over HTTP. Which deployment type satisfies this requirement?

  1. A catalog asset published to the glossary
  2. A notebook run in interactive mode
  3. A batch deployment triggered by a schedule
  4. An online deployment exposing a scoring endpoint Correct answer.

Explanation

Online deployments expose a REST endpoint that applications call to receive predictions in real time. Batch deployments score stored data on a schedule and do not offer per-request serving.

Question 18Exam objectives (not enumerated in retrieved source)

Which factor most directly determines whether a Cloud Pak for Data installation can run on a customer's existing infrastructure?

  1. Whether the customer has a mainframe in the data center
  2. Whether the customer uses spreadsheets for reporting
  3. Whether the infrastructure provides a supported Red Hat OpenShift cluster with adequate resources Correct answer.
  4. Whether the customer's staff prefer Python over R

Explanation

The platform is delivered as operators on Red Hat OpenShift, so a supported cluster with sufficient compute and storage is the gating requirement. Language preferences and reporting habits do not affect installability.

Question 19Exam objectives (not enumerated in retrieved source)

What does a data quality assessment in Cloud Pak for Data typically produce for a dataset?

  1. A new OpenShift node dedicated to the dataset
  2. A permanent lock preventing further edits
  3. A replacement glossary term for every column
  4. A set of metrics and findings describing completeness, validity, and other dimensions Correct answer.

Explanation

Quality assessments analyze data against defined dimensions and report metrics and issues for review. The output guides remediation rather than provisioning infrastructure or locking assets.

Question 20Exam objectives (not enumerated in retrieved source)

An analyst must ensure that sensitive columns are masked when other users query a virtualized table. Which combination of platform features applies?

  1. Publishing the asset to a public catalog without restrictions
  2. Adding more replicas to the Db2 warehouse
  3. Data protection rules combined with governance policies on the asset Correct answer.
  4. Increasing the query timeout on the virtualization engine

Explanation

Data protection rules identify sensitive columns and apply masking or denial based on policy, while governance records document the classification. Together they enforce masking for unauthorized consumers.

Question 21Exam objectives (not enumerated in retrieved source)

Which statement best explains why metadata import matters when onboarding a new database into the catalog?

  1. It encrypts the source database automatically
  2. It registers technical metadata such as schemas and columns so assets become discoverable and governable Correct answer.
  3. It copies all table rows into the catalog for safekeeping
  4. It converts the database into a machine learning model

Explanation

Metadata import harvests structural information about tables and columns, creating catalog entries that users can search and stewards can enrich. The data itself stays in the source system.

Question 22Exam objectives (not enumerated in retrieved source)

A team notices that a deployed model's predictions are not explainable to auditors. Which capability should they enable?

  1. Automatic glossary term creation
  2. Model explainability features that generate feature-attribution insights Correct answer.
  3. Additional catalog tags on the training data
  4. A larger notebook runtime

Explanation

Explainability generates per-prediction or global feature attributions that show which inputs influenced the output. This documentation supports audit and regulatory review of model decisions.

Question 23Exam objectives (not enumerated in retrieved source)

Which consideration is most important when choosing between a batch and an online deployment for a scoring workload?

  1. Whether the model was written in a notebook
  2. Whether the catalog uses glossary terms
  3. Whether predictions must be returned per request in real time or can be computed on stored data periodically Correct answer.
  4. Whether the project has a Git repository attached

Explanation

The decisive factor is the latency and access pattern of consumers: request-time scoring needs an online endpoint, while scheduled scoring of stored records fits batch. Authoring details do not determine the serving pattern.

Question 24Exam objectives (not enumerated in retrieved source)

What is the purpose of associating a connection asset with a project?

  1. To define the business definition of a data element
  2. To allocate CPU resources to the analytics engine
  3. To store the endpoint and credentials needed to access a data source Correct answer.
  4. To schedule model retraining automatically

Explanation

A connection asset holds the endpoint, authentication, and driver details required to reach a source. Projects reference connections so users can query the source without re-entering credentials.

Question 25Exam objectives (not enumerated in retrieved source)

An organization wants to compare a model's predictions against actual outcomes recorded later in a feedback table. Which OpenScale concept supports this?

  1. The project's file upload limit
  2. The cluster's ingress configuration
  3. Feedback logging linked to the monitored deployment Correct answer.
  4. The catalog's publish workflow

Explanation

Feedback logging records actual outcomes and associates them with prior predictions so accuracy and other quality metrics can be computed. Without feedback, monitoring can only observe payload characteristics.

Question 26Exam objectives (not enumerated in retrieved source)

Which Cloud Pak for Data persona is primarily responsible for curating and governing data assets in the catalog?

  1. Data engineer
  2. Data steward Correct answer.
  3. Data scientist
  4. Business analyst

Explanation

The data steward persona focuses on catalog curation, glossary management, and enforcing data governance policies. Data engineers build pipelines, data scientists create models, and business analysts consume data for insights, but stewardship is the governance-centric role.

Question 27Exam objectives (not enumerated in retrieved source)

What is the main benefit of using a virtualized asset in a Cloud Pak for Data project instead of copying data into the project?

  1. It stores a full physical copy for faster local processing.
  2. It queries the source system in place, avoiding data duplication and staleness. Correct answer.
  3. It automatically cleanses the data during query execution.
  4. It encrypts the data at rest in the remote source.

Explanation

Virtualized assets let users query remote data through a connection without importing it, so the data stays in its source system and remains current. This reduces storage duplication and avoids stale copies, but it does not cleanse data or encrypt the source.

Question 28Exam objectives (not enumerated in retrieved source)

Which Cloud Pak for Data component provides a unified SQL query engine that can join data across multiple federated sources?

  1. DataStage
  2. Db2 Warehouse
  3. Watson Studio
  4. Data Virtualization Correct answer.

Explanation

Data Virtualization provides a federated SQL engine that presents remote tables as virtual tables and can join them across sources. Watson Studio is for model building, DataStage for ETL, and Db2 Warehouse is a target warehouse, not a federation layer.

Question 29Exam objectives (not enumerated in retrieved source)

In Cloud Pak for Data, which asset type is used to store the connection details needed to access an external database?

  1. Data set
  2. Job
  3. Connection Correct answer.
  4. Notebook

Explanation

A connection asset holds host, port, credentials, and driver information required to reach an external system. Data sets hold files, notebooks hold code, and jobs schedule tasks; none of these store connection properties.

Question 30Exam objectives (not enumerated in retrieved source)

What does a governance artifact called a business term represent in the Cloud Pak for Data catalog?

  1. A machine learning model registered for deployment.
  2. A physical table stored in the data lake.
  3. A controlled vocabulary definition that standardizes business concepts across assets. Correct answer.
  4. A scheduled job that refreshes metadata.

Explanation

Business terms are governed vocabulary entries that give shared meaning to data assets so different teams interpret concepts consistently. They are not physical tables, scheduled jobs, or models; they are metadata definitions managed by stewards.

Question 31Exam objectives (not enumerated in retrieved source)

Which Cloud Pak for Data service is designed to profile, cleanse, and monitor data quality across sources?

  1. Data Refinery Correct answer.
  2. Watson Studio
  3. Data Virtualization
  4. Watson Knowledge Catalog

Explanation

Data Refinery provides visual data preparation, profiling, and cleansing operations that can be saved and rerun. Knowledge Catalog governs metadata, Watson Studio builds models, and Data Virtualization federates queries; none of these is the primary cleansing tool.

Question 32Exam objectives (not enumerated in retrieved source)

A data engineer needs to schedule a recurring transformation job in a Cloud Pak for Data project. Which capability should be used?

  1. A notebook run manually each time
  2. A deployment space
  3. A catalog glossary term
  4. A job created from the project asset with a schedule Correct answer.

Explanation

Jobs in a project can be created from notebooks, Data Refinery flows, or other assets and given a schedule so they run automatically. Manual notebook runs, glossary terms, and deployment spaces do not provide recurring scheduling.

Question 33Exam objectives (not enumerated in retrieved source)

Which statement describes the purpose of a catalog in Cloud Pak for Data?

  1. It executes machine learning training jobs on GPUs.
  2. It defines the network topology for the cluster.
  3. It stores raw files for processing by notebooks.
  4. It provides a searchable inventory of governed data assets with metadata and lineage. Correct answer.

Explanation

A catalog is a governed inventory that makes assets discoverable through metadata, classifications, and lineage. It does not store raw files for processing, run training jobs, or define network topology; those belong to storage, compute, and infrastructure layers.

Question 34Exam objectives (not enumerated in retrieved source)

A data scientist wants to reuse a Python function across multiple notebooks in the same project. What is the recommended approach in Cloud Pak for Data?

  1. Embed the function in a catalog glossary term.
  2. Save the function as a deployment space asset.
  3. Store the function in a project Git repository and import it. Correct answer.
  4. Copy the function into every notebook manually.

Explanation

A Git repository integrated with the project allows code to be versioned and imported into notebooks, promoting reuse and consistency. Manual copying causes drift, and glossary terms or deployment spaces are not code repositories.

Question 35Exam objectives (not enumerated in retrieved source)

Which access control mechanism in Cloud Pak for Data determines what a user can do with a specific project asset?

  1. Asset-level permissions combined with project roles Correct answer.
  2. Network firewall rules
  3. Platform role assignments only
  4. Catalog metadata import schedules

Explanation

Access to project assets is governed by a combination of project roles and asset-level permissions, allowing fine-grained control. Platform roles set broad capabilities, while firewalls and import schedules do not govern asset actions.

Question 36Exam objectives (not enumerated in retrieved source)

An analyst wants to preview the first 100 rows of a virtualized table without writing SQL. Which Cloud Pak for Data feature supports this?

  1. The catalog lineage graph
  2. The data asset preview in the project Correct answer.
  3. The glossary term editor
  4. The deployment space monitoring dashboard

Explanation

The data asset preview in a project displays a sample of rows so users can inspect content quickly. Lineage graphs show relationships, monitoring dashboards show deployment metrics, and glossary editors manage definitions; none of these previews rows.

Question 37Exam objectives (not enumerated in retrieved source)

Which Cloud Pak for Data capability helps a steward understand the impact of changing a source column on downstream assets?

  1. Data replication
  2. Job scheduling
  3. Data lineage Correct answer.
  4. Model drift detection

Explanation

Data lineage records how data moves and transforms across assets, so a steward can trace downstream dependencies before making changes. Replication copies data, drift detection monitors models, and scheduling automates jobs; none of these show impact relationships.

Question 38Exam objectives (not enumerated in retrieved source)

A team needs to share a trained model with another team for scoring without exposing the training data. Which Cloud Pak for Data construct is appropriate?

  1. A deployment space with the model deployed Correct answer.
  2. A catalog containing the raw training data
  3. A project containing the training notebook
  4. A glossary term describing the model

Explanation

A deployment space hosts deployed models and provides scoring endpoints, allowing other teams to consume the model without access to training data. Projects, catalogs of raw data, and glossary terms do not provide isolated scoring access.

Question 39Exam objectives (not enumerated in retrieved source)

Which metadata import method is suitable for continuously synchronizing technical metadata from a relational database into the catalog?

  1. A one-time manual metadata import
  2. A notebook that prints table names
  3. A scheduled metadata import job Correct answer.
  4. A data refinery flow

Explanation

Scheduled metadata import jobs can run repeatedly to keep catalog metadata aligned with source changes. A one-time import becomes stale, while Data Refinery and notebooks do not update catalog metadata automatically.

Question 40Exam objectives (not enumerated in retrieved source)

What is the effect of assigning a user the Viewer role on a catalog?

  1. The user can edit glossary terms and approve assets.
  2. The user can view and search catalog assets but cannot modify them. Correct answer.
  3. The user can manage platform users and roles.
  4. The user can delete catalog assets.

Explanation

The Viewer role grants read and search access to catalog content without modification rights. Editing glossary terms, deleting assets, and managing users require higher-privilege roles such as Editor or Administrator.

Question 41Exam objectives (not enumerated in retrieved source)

A data engineer must ensure that a Data Refinery flow runs only after an upstream table has been refreshed. Which Cloud Pak for Data feature supports this dependency?

  1. Manual execution by the engineer
  2. A glossary workflow
  3. Job scheduling with dependency on another job Correct answer.
  4. A catalog asset rating

Explanation

Job scheduling can define dependencies so that a downstream job starts only after an upstream job completes successfully. Manual execution is unreliable, and glossary workflows or asset ratings do not control job execution order.

Question 42Exam objectives (not enumerated in retrieved source)

Which statement best describes how Cloud Pak for Data handles sensitive data when a user queries a virtualized table with a masking policy applied?

  1. The masking is applied only when data is exported to a file.
  2. The masking requires the user to manually edit the SQL.
  3. The masking is enforced at query time so the user sees masked values. Correct answer.
  4. The masking is applied only to catalog metadata, not query results.

Explanation

Masking policies are enforced at query time so that users without privileged access receive masked values in result sets. The masking is not limited to exports or metadata, and users do not need to rewrite SQL to benefit from it.

Question 43Exam objectives (not enumerated in retrieved source)

A team wants to document the meaning of a newly onboarded column and link it to a governed definition. Which two catalog artifacts should they use together?

  1. A deployment space and a model
  2. A data set and a notebook
  3. A business term and an asset assignment Correct answer.
  4. A job and a schedule

Explanation

A business term provides the governed definition, and assigning it to the column asset links the technical metadata to that meaning. Data sets, notebooks, deployment spaces, models, and jobs do not establish governed semantic definitions.

Question 44Exam objectives (not enumerated in retrieved source)

Which factor most affects the choice between using a notebook and using Data Refinery for a transformation task in Cloud Pak for Data?

  1. The physical location of the data center
  2. Whether the transformation requires custom code or visual, repeatable steps Correct answer.
  3. The color theme of the user interface
  4. The number of users in the catalog

Explanation

Notebooks suit custom code and complex logic, while Data Refinery suits visual, repeatable transformations that can be scheduled. Catalog user counts, interface themes, and data center location do not determine the transformation tool choice.

Question 45Exam objectives (not enumerated in retrieved source)

An administrator needs to give a group of users the ability to create projects but not to manage platform settings. Which approach is correct?

  1. Assign the Administrator platform role to each user.
  2. Add each user to every project as an Editor.
  3. Assign a platform role that includes project creation without administrative privileges. Correct answer.
  4. Share the administrator password with the group.

Explanation

Platform roles define capabilities such as project creation; assigning a non-administrative role that includes project creation grants the needed ability without platform-wide control. Administrator roles grant excessive privileges, project Editor roles do not enable project creation, and sharing passwords is insecure.

Question 46Exam objectives (not enumerated in retrieved source)

A data steward wants to ensure that a term used in a report matches the approved definition. Which catalog feature should be used?

  1. A notebook
  2. A governance workflow for term approval Correct answer.
  3. A data quality rule
  4. A deployment space

Explanation

Governance workflows route business terms through review and approval so that only sanctioned definitions are published. Data quality rules check data values, deployment spaces host models, and notebooks run code; none of these approve terminology.

Question 47Exam objectives (not enumerated in retrieved source)

Which statement accurately describes how a project interacts with a catalog when a user publishes an asset?

  1. Publishing deletes the asset from the project.
  2. Publishing encrypts the asset in the project.
  3. Publishing copies the asset into the catalog as a governed, discoverable entry. Correct answer.
  4. Publishing moves the asset to a deployment space.

Explanation

Publishing an asset from a project adds it to the catalog so it can be discovered, governed, and reused by others. The project asset remains, and publishing does not move assets to deployment spaces or encrypt them.

Question 48Exam objectives (not enumerated in retrieved source)

A team needs to compare two versions of a data set to identify which rows changed between refreshes. Which Cloud Pak for Data capability supports this comparison?

  1. Data profiling and comparison in Data Refinery Correct answer.
  2. Deployment space monitoring
  3. Data lineage visualization
  4. Glossary term assignment

Explanation

Data Refinery profiling and comparison operations can highlight differences between data set versions, such as changed or added rows. Lineage shows relationships, glossary terms define meaning, and deployment monitoring tracks models, not row-level differences.

Question 49Exam objectives (not enumerated in retrieved source)

An organization must ensure that only approved models can be promoted to production in a deployment space. Which governance control should be implemented?

  1. Disable model monitoring to reduce overhead.
  2. Allow all project members to deploy models directly.
  3. Store all models in a shared notebook.
  4. Require model approval through a governed workflow before deployment. Correct answer.

Explanation

A governed approval workflow ensures that only reviewed and sanctioned models are promoted to production, supporting compliance and quality. Open deployment rights, disabled monitoring, and storing models in notebooks do not enforce approval controls.

Question 50Exam objectives (not enumerated in retrieved source)

Which combination best supports end-to-end traceability from a source column to a deployed model's input in Cloud Pak for Data?

  1. A glossary term plus a notebook
  2. Data lineage plus model metadata registration Correct answer.
  3. A job schedule plus a catalog rating
  4. A deployment space plus a data set

Explanation

Data lineage traces the column through transformations, and model metadata registration records which inputs a model uses, together providing end-to-end traceability. Glossary terms, notebooks, schedules, ratings, deployment spaces, and data sets alone do not link source columns to model inputs.