Free Salesforce Platform Data Architect practice questions

Question 1

At Ursa Major Solar, the custom Site Visit object has a Private organization-wide default. Two field engineers who hold the same Field Engineer West role report that they can't see each other's site visits, although their regional manager sees both. Engineers in that role must see all of each other's visits, and no one else's access should change. What should the Salesforce architect recommend?
A. Move one of the engineers into a subordinate role, so the other engineer inherits access through the hierarchy.
B. Change the Site Visit default to Public Read Only, so engineers in the same role can see each other's visits.
C. Have each engineer manually share every visit with the peer, because sharing rules can't target one role.
D. Add an owner-based sharing rule that shares visits owned by Field Engineer West with that same role.

Accepted Answer

Correct answer: D. Add an owner-based sharing rule that shares visits owned by Field Engineer West with that same role.
Users in the same role don't automatically get access to each other's records, and Salesforce architect guidance lists giving peers who hold the same role access as a use case for owner-based sharing rules. Moving one engineer down a level would grant access in only one direction and distort the hierarchy. A Public Read Only default would change everyone's access to site visits, not just the engineers'. Sharing rules can share records owned by a role with that same role, and manual shares don't scale and are removed when the owner changes.

Question 2

Bloomington Caregivers plans to store about 300 million home-visit notes. The data architect is collecting requirements before choosing between a custom object and a custom big object. Which two requirements mean the visit notes should stay in a custom object rather than a big object? Choose 2 answers.
A. Coordinators must see only the visits for patients in their own region, enforced through sharing rules on each visit record.
B. Notes must be kept for 15 years so that regulators can review the full visit history of any patient long after care ends.
C. A record-triggered flow must create a follow-up task when a note records a missed medication dose during the visit.
D. Most lookups read notes for a single patient within a date range, using the patient and the visit date as filters.
E. The notes table is expected to grow past several hundred million rows within the next few years of operation.

Accepted Answer

Correct answer: A. Coordinators must see only the visits for patients in their own region, enforced through sharing rules on each visit record. | C. A record-triggered flow must create a follow-up task when a note records a missed medication dose during the visit.
Big objects support only object and field permissions, not regular or standard sharing rules, so region-based row visibility needs a custom object, and features like triggers and flows aren't supported on big objects, so a record-triggered flow on new notes rules them out too. Long retention is a reason to choose a big object, which Salesforce positions for historical archives. Reads by patient and date map naturally to a big object index. Growth to hundreds of millions of rows is exactly the scale big objects are designed for.

Question 3

Cumulus Financial links its custom Advisor and Portfolio objects many-to-many through a new Advisor Assignment junction object that has a master-detail relationship to each of them. Each assignment stores the advisor's role on the portfolio and a start date. Which two statements should the data architect include in the design review? Choose 2 answers.
A. Roll-up summary fields that summarize Advisor Assignment records can be created on both the Portfolio and the Advisor objects.
B. Advisor Assignment records keep an owner of their own, so new assignments can be routed to a queue for approval.
C. Advisor Assignment can later become the master of a custom Assignment Note object to hold meeting notes per assignment.
D. Formula fields and validation rules on Advisor Assignment can reference fields on both the Portfolio and the Advisor.
E. Deleting an Advisor record leaves that advisor's assignment records in place with an empty Advisor field for audit.

Accepted Answer

Correct answer: A. Roll-up summary fields that summarize Advisor Assignment records can be created on both the Portfolio and the Advisor objects. | D. Formula fields and validation rules on Advisor Assignment can reference fields on both the Portfolio and the Advisor.
Salesforce lets you build roll-up summary fields on both master objects of a junction, and formula fields and validation rules on the junction can reference fields on both masters. Junction records have no owner of their own: they inherit the Owner value from the primary master record, and detail-side objects can't use queues. A junction object can't be on the master side of another master-detail relationship, so it can't take a detail object such as Assignment Note. Junction records are deleted when either master record is deleted, so they aren't kept with an empty master field.

Question 4

Universal Containers is opening a subsidiary in Brazil. The compliance officer wants stewards to tag fields that hold data covered by Brazil's LGPD privacy law using the same field-level classification they already use for GDPR and CCPA, so reports on regulated fields stay consistent. What should the data architect recommend?
A. Add an LGPD value to the Data Sensitivity Level picklist, because that picklist lists the privacy laws that apply to each field.
B. Add an LGPD value through Edit Compliance Categorization Picklist Values in Data Classification Settings, then tag the fields.
C. Create a custom Regulation__c picklist field on every object, because the Compliance Categorization values are fixed by Salesforce.
D. Open a case with Salesforce Support to add LGPD, because only Support can change the default compliance categorization values.

Accepted Answer

Correct answer: B. Add an LGPD value through Edit Compliance Categorization Picklist Values in Data Classification Settings, then tag the fields.
Compliance Categorization records the acts or regulations related to a field's data, and its values can be customized by selecting Edit Compliance Categorization Picklist Values on the Data Classification Settings page (or by updating the ComplianceGroup StandardValueSet), so adding LGPD keeps one consistent scheme. Data Sensitivity Level records how sensitive the data is, not which law applies. The compliance values aren't fixed, so a custom Regulation__c field on every object would duplicate built-in metadata at record level. Admins can edit the values themselves, so Support isn't needed.

Question 5

Bloomington Caregivers' data steward is consolidating client accounts into one trusted record per family in Lightning Experience. For most duplicates she clicks View Duplicates and merges them. Two accounts for the same family, Kim Household and The Kim Family, aren't flagged by the active duplicate rule, so no merge option appears for them. She has Delete permission on accounts and is a Salesforce admin. What should the data architect recommend?
A. Merge the pair with the Salesforce Classic merge tool, and tune the matching rule so that similar duplicates are flagged in future
B. Grant the steward the Modify All Data permission, because the merge option appears only for users who can modify every account
C. Delete one account and reparent its contacts by hand, because accounts that a rule doesn't flag can never be merged
D. Change the duplicate rule action to Block, because Lightning Experience offers merges only for duplicates the rule blocks

Accepted Answer

Correct answer: A. Merge the pair with the Salesforce Classic merge tool, and tune the matching rule so that similar duplicates are flagged in future
Salesforce Help's Lightning Experience merge article says duplicates are merged from the View Duplicates option, that admins can fine-tune matching rules when duplicates aren't flagged, and that accounts not detected by duplicate rules can be merged with the merge tool in Salesforce Classic. The steward already meets the merge permissions, Delete on accounts plus being an admin, the owner or above the owner in the role hierarchy, so Modify All Data isn't what's missing. Deleting and reparenting by hand loses what a merge preserves and isn't necessary. The rule's Allow or Block action decides what happens when a duplicate is saved; it doesn't make an unmatched pair appear as duplicates.

Question 6

In a current Data 360 org: Cumulus Financial runs separate Data 360 (formerly Data Cloud) instances for its retail and insurance businesses because of regulatory separation. The retail instance shares its High-Value Customer segment and two data model objects with the insurance instance through a data share. In the insurance instance, the governance policies that restrict access to data tagged as PII don't apply to the shared objects, although the same fields are tagged in the retail instance. What explains this?
A. Tags aren't shared with data shares, so the insurance instance must reapply its own tags before its tag-based policies can apply
B. The data share copies the objects into the insurance instance overnight, and the tags arrive only after the next full refresh runs
C. Data shares work only for segments, so the shared data model objects bypass governance until they're shared again as segments
D. Tag-based policies in the receiving instance always defer to the source instance's policies, so they are switched off by design

Accepted Answer

Correct answer: A. Tags aren't shared with data shares, so the insurance instance must reapply its own tags before its tag-based policies can apply
The provisioning guide states that tags aren't shared between Data 360 orgs: only the raw objects are made available, and the target org must reapply any governance, operational or AI tags. Data sharing is zero copy, so nothing is copied overnight and there's no refresh that later delivers tags. Data shares can include data model objects and calculated insights as well as segments. Nothing in the guidance says receiving-org policies defer to the source org; governance is federated, so the receiving org applies its own controls once its tags exist.

Question 7

AW Computing has 120 billing specialists on Salesforce Platform licenses who work in a custom invoicing app. The company now wants these same specialists to act as support agents, working customer cases from a queue and resolving them against entitlements. The admin plans to add Case create and edit through a permission set. What should the Salesforce architect advise?
A. Proceed with the permission set, because Salesforce Platform licenses include Case create and edit, which is all agent case work requires
B. Assign Customer Community Plus licenses to the specialists, because case work for customers needs a customer-facing license
C. Assign Knowledge Only User licenses, because agents mainly need Knowledge articles to resolve customer cases
D. Provide Service Cloud licenses for the specialists who will act as support agents, as Platform case access doesn't cover agent work

Accepted Answer

Correct answer: D. Provide Service Cloud licenses for the specialists who will act as support agents, as Platform case access doesn't cover agent work
The Lightning Platform license details say Platform-licensed users must be internal employees or contractors and can't complete internal or external customer service work without a Service Cloud license, and the Cases row notes that a Service Cloud license is required for users acting as support agents. The permission set only grants object access the license already exposes; it doesn't change what the license entitles. Customer Community Plus is an external license that must not be assigned to employees. The Knowledge Only User license is limited to the Knowledge app, custom objects and a few standard tabs, so it doesn't support agent case work.

Question 8

Cumulus Financial's rate engine sends loan interest rates with four decimal places, such as 5.2375, to Rate__c, a custom Percent field on Loan__c whose Decimal Places attribute is 2. The integration logs no errors, but the nightly reconciliation finds stored values such as 5.24 that don't match the engine. What should the data architect recommend?
A. Keep two decimal places, because the API cuts off the extra digits so that 5.2375 is stored as 5.23 and the engine can do the same
B. Change Rate__c to a Number field, because Percent fields always store exactly two decimal places whatever their attribute says
C. Select Allow field truncation in the integration's settings, so that the extra digits are kept in the field instead of being rounded
D. Set Rate__c's Decimal Places to 4 and reload the rates, because the platform rounds entered decimals to the field's decimal places

Accepted Answer

Correct answer: D. Set Rate__c's Decimal Places to 4 and reload the rates, because the platform rounds entered decimals to the field's decimal places
Salesforce Help's Custom Field Attributes page says Decimal Places sets how many digits can be entered to the right of the decimal point for currency, geolocation, number and percent fields, and that the system rounds the decimal numbers entered when necessary, so 5.2375 becomes 5.24 in a two-decimal field and the field needs four decimal places to keep the engine's precision. The value is rounded, not cut off, which is why 5.24 appears instead of 5.23. Percent fields have a configurable Decimal Places attribute, so changing the type isn't needed. Field truncation is a text-length setting and doesn't add decimal precision to a number.

Question 9

Cloud Kicks' online store created a person account for the shopper Dana Ortiz. Years earlier, its retail point-of-sale integration had created a business account named Dana Ortiz for the same person. The data steward wants one record for her, but the merge screen won't let her select both accounts. What should the data architect recommend?
A. Change the business account's record type to a person account record type through the API, and then merge the two person accounts
B. Use the merge tool in Salesforce Classic, which can merge a person account with a business account when duplicates aren't flagged
C. Relate the person account to the business account with Contacts to Multiple Accounts so both records show the shared customer
D. Select the business account as the principal record on the merge screen, which lets the merge accept both accounts

Accepted Answer

Correct answer: A. Change the business account's record type to a person account record type through the API, and then merge the two person accounts
Salesforce Help says person accounts can be merged only with other person accounts, and that changing an account's record type from a business account to a person account requires the API, so the business account must first become a person account before the two records can be merged. Help's merge considerations also say that business and person accounts can't be merged with each other in Salesforce Classic either. Contacts to Multiple Accounts links the two records but keeps both, so the duplicate remains. Choosing the business account as the principal record doesn't change the rule that a person account can't be merged with a business account.

Question 10

Cloud Kicks encrypted a custom Passport_Number__c field on Contact with Shield Platform Encryption. While troubleshooting a ticket, a system administrator logs in as a sales rep and sees the passport number in plain text. The privacy officer assumed encryption would hide the value from everyone except the HR team and asks the data architect to explain. What is the cause?
A. The sales rep's profile has field-level security access to the field, and encryption at rest doesn't change who can see decrypted values
B. The deterministic scheme was chosen for the field, and deterministic values are shown decrypted to every user so that they can be filtered
C. Administrators who log in as other users bypass Shield Platform Encryption, so the value would be masked if the rep viewed it directly
D. The tenant secret was rotated after the field was encrypted, so the existing data stayed in plain text until it is synchronised again

Accepted Answer

Correct answer: A. The sales rep's profile has field-level security access to the field, and encryption at rest doesn't change who can see decrypted values
Shield Platform Encryption isn't an authentication or authorization tool; users with field-level security access to an encrypted field see its plaintext, including someone using Login As for that user, so access must be restricted with field-level security, page layouts and sharing. Both deterministic and probabilistic schemes encrypt data at rest and neither controls who sees decrypted values. Login As works within the target user's own permissions rather than bypassing encryption, so the rep sees the same value directly. Rotating a tenant secret leaves existing data encrypted under the earlier key rather than turning it back into plain text.

Question 11

Ursa Major Solar is writing its first data management strategy before migrating 12 million service records from three legacy systems into Salesforce. The data architect drafts the questions the strategy document must answer. Which three questions should the architect include? Choose 3 answers.
A. How much data will be stored, and what is the potential growth rate of that data over the coming years?
B. How will record duplication be managed and resolved as data arrives from the three legacy systems?
C. What is the delete and archive strategy for redundant data that is no longer needed by the business?
D. Which Lightning page templates and record page themes will each user group see when the migration goes live?
E. Which sandbox naming conventions will each development team use while it builds the migration scripts?

Accepted Answer

Correct answer: A. How much data will be stored, and what is the potential growth rate of that data over the coming years? | B. How will record duplication be managed and resolved as data arrives from the three legacy systems? | C. What is the delete and archive strategy for redundant data that is no longer needed by the business?
Trailhead's data management guidance lists questions a strategy should answer at a minimum, including how much data you will store and its growth rate, how you will manage and resolve duplication, and what the delete or archive strategy is for data that is no longer needed. Lightning page templates and themes are user experience design choices, not data management policy. Sandbox naming conventions are a release management detail and don't shape how data is captured, secured or retired. The three data questions drive storage, quality and retention decisions that the governance program must own.

Question 12

Get Cloudy Consulting loads ERP data into a custom staging object, Staging_Load__c, which now holds 30 million processed rows. The data architect wants every row removed permanently while keeping the object, its fields and its triggers for future loads. Truncation is enabled in User Interface settings, but Salesforce won't truncate the object. Staging_Load__c has a Legacy_Key__c text field marked as External ID, and no other object looks it up. What should the data architect do?
A. Delete the rows in Bulk API 2.0 batches until fewer than 10 million remain, because truncation is blocked on custom objects above that record count.
B. Deactivate the object's triggers and validation rules first, because truncation would otherwise delete them along with the records it removes.
C. Empty the object's Recycle Bin of rows from an earlier delete, because truncation can't run while deleted rows of the object are still recoverable.
D. Remove the External ID attribute and any custom index from Legacy_Key__c, truncate the object, then add them back.

Accepted Answer

Correct answer: D. Remove the External ID attribute and any custom index from Legacy_Key__c, truncate the object, then add them back.
Salesforce Help lists objects that can't be truncated, including custom objects that have a custom index or an external ID, so the External ID on Legacy_Key__c blocks truncation until it and any custom index are removed. There is no record-count threshold for truncation. Truncation preserves the object definition, triggers, validation rules and sharing rules, so they don't need to be deactivated to survive. Truncation itself erases the records currently in the object's Recycle Bin, so an earlier delete doesn't block it.

Question 13

Universal Containers wants two kinds of external data in Salesforce without copying them in: signed contract PDFs stored in a cloud content system, which users should browse and attach from Salesforce, and order rows from an SAP gateway that publishes an OData 4.0 service. The org runs Enterprise Edition with the Salesforce Connect add-on. Which two options should the data architect recommend? Choose 2 answers.
A. Files Connect with an external data source for the cloud content system, so users reach the contract files from inside Salesforce.
B. Salesforce Connect with the OData 4.0 adapter, so the SAP order rows appear as external objects that are read on demand.
C. Data Loader jobs that upload the contract PDFs into Salesforce Files each night, so the files are searchable inside the org.
D. The Salesforce Connect cross-org adapter pointed at the SAP gateway, because it reads any REST endpoint without extra setup.
E. A custom big object that stores the SAP order rows, so the high volume doesn't count against the org's data storage.

Accepted Answer

Correct answer: A. Files Connect with an external data source for the cloud content system, so users reach the contract files from inside Salesforce. | B. Salesforce Connect with the OData 4.0 adapter, so the SAP order rows appear as external objects that are read on demand.
Both Files Connect and Salesforce Connect use external data sources: Files Connect reaches third-party content systems, and Salesforce Connect maps external tables to external objects, so the contracts and the orders stay in their source systems. The OData 4.0 adapter fits an SAP gateway that publishes an OData producer. Nightly Data Loader uploads copy the PDFs into Salesforce, which the requirement rules out. The cross-org adapter uses the Lightning Platform REST API to reach other Salesforce orgs, not arbitrary REST endpoints. A big object stores the rows in Salesforce, so it isn't virtualization.

Question 14

Ursa Major Solar dispatchers use list views on an 18 million row Work_Visit__c object. Each list view selects visits for one Service_Territory__c value and sorts them by Visit_Date__c. Both fields are custom, and the territory filter is selective. What should the data architect recommend to make these list views faster?
A. Ask for a two-column custom index with Service_Territory__c in the first column and Visit_Date__c in the second column.
B. Ask for a two-column custom index with Visit_Date__c in the first column and Service_Territory__c in the second column.
C. Ask for two separate single-column custom indexes, one on Service_Territory__c and one on Visit_Date__c, for each list view.
D. Enable divisions and create one division per service territory, so each list view only reads rows from the dispatcher's division.

Accepted Answer

Correct answer: A. Ask for a two-column custom index with Service_Territory__c in the first column and Visit_Date__c in the second column.
Two-column custom indexes are useful for list views that use one field to select records and another to sort them, with the selecting field first; for a filter plus sort pattern they are more efficient than separate single indexes. Putting Visit_Date__c first reverses the pattern, so the index wouldn't match the territory selection. Two single-column indexes serve the filter but not the combined select-and-sort as efficiently. Divisions partition large deployments into broad groups such as regions and are enabled by Support; they aren't a per-territory sorting tool.

Question 15

Northern Trail Outfitters has just defined an external data source for Salesforce Connect with the OData 4.0 adapter and synced it to create the external object Trail_Permit__x. In list views, testers see the same permit twice on some pages, and sorting by Issue Date returns rows in an odd order. What is the best next step for the data architect?
A. Increase the external data source's page size, so every permit loads in a single page and duplicates can't appear.
B. Run the validator tool on Trail_Permit__x, which tests the connection for ID uniqueness and the ability to sort and filter results.
C. Delete the external object and recreate it manually instead of syncing, because synced external objects can't sort on date columns.
D. Add a unique External ID custom field to Trail_Permit__x, so Salesforce rejects duplicate permit rows when the list view loads them.

Accepted Answer

Correct answer: B. Run the validator tool on Trail_Permit__x, which tests the connection for ID uniqueness and the ability to sort and filter results.
Salesforce recommends running the validator tool on each external object after configuring the external data source; it tests for ID uniqueness and for the ability to sort and filter, which are the symptoms the testers see. A larger page size changes how many rows each callout returns but doesn't fix duplicate IDs or broken sorting in the producer. Synced external objects are the recommended approach and can sort on date columns when the producer supports it. External objects don't store the rows, so a unique field in Salesforce can't reject duplicates returned by the external system.

Question 16

Get Cloudy Consulting is loading 6 million custom Engagement__c records with Data Loader, with Use Bulk API 2.0 selected. Allow field truncation is selected in the settings, yet about 12,000 rows fail because their legacy Summary__c values are longer than the 255-character text field. What is the cause, and what should the data architect recommend?
A. Bulk API 2.0 truncates oversized values only on update, so the architect should insert the rows first and then update the summaries in a second job
B. Field truncation applies only to Long Text Area fields, so the architect should convert Summary__c to Long Text Area before the next load
C. Allow field truncation isn't available with the Bulk APIs, so those rows fail; the architect should resolve the long values in staging or widen the field
D. Bulk API 2.0 rejects oversized rows until the job finishes, so the architect should resubmit the same file to let the truncation take effect

Accepted Answer

Correct answer: C. Allow field truncation isn't available with the Bulk APIs, so those rows fail; the architect should resolve the long values in staging or widen the field
The Data Loader guide says Allow field truncation isn't available when Use Bulk API or Use Bulk API 2.0 is selected, and in that case a row fails when a value is too large for the field. The quality decision belongs in staging: either trim or summarise the text deliberately, or change the target field to a type that holds the full value. Bulk API 2.0 doesn't truncate on update either, so a second job fails the same way. Truncation covers text, email, phone and picklist types, not only long text fields. Resubmitting the same file reproduces the same failures because nothing about the data has changed.

Question 17

Universal Containers extracts its 60-million-row Order_Line__c object nightly to a data lake with this Bulk API 2.0 query job:

SELECT Id, Order__c, SKU__c, Amount__c
FROM Order_Line__c
WHERE LastModifiedDate = LAST_N_DAYS:2
ORDER BY CreatedDate
LIMIT 40000000

The job keeps timing out, and the data lake sorts and deduplicates rows on arrival. Which two changes should the data architect recommend? Choose 2 answers.
A. Remove the ORDER BY CreatedDate clause, because an ORDER BY clause disables PK chunking for the query job.
B. Add the Sforce-Enable-PKChunking: chunkSize=250000 header to the job creation request to force smaller chunks.
C. Change the job's operation from query to queryAll so that Salesforce can use the recycle bin index.
D. Remove the LIMIT 40000000 clause, because a LIMIT clause disables PK chunking and leads to query timeouts.
E. Lower the maxRecords value on the job creation request so that each internal chunk processes fewer rows.

Accepted Answer

Correct answer: A. Remove the ORDER BY CreatedDate clause, because an ORDER BY clause disables PK chunking for the query job. | D. Remove the LIMIT 40000000 clause, because a LIMIT clause disables PK chunking and leads to query timeouts.
Bulk API 2.0 chunks large query jobs automatically, but LIMIT and ORDER BY disable PK chunking, which makes queries take longer and can cause timeouts; Salesforce recommends removing them before further troubleshooting, and the data lake already sorts and deduplicates rows. The Sforce-Enable-PKChunking header is how the original Bulk API enables chunking manually, while Bulk API 2.0 decides chunking itself. The queryAll operation adds deleted and archived records and has no performance benefit. maxRecords is a parameter of the Get Results request that limits the size of each downloaded result set, not the way the query is chunked.

Question 18

During a migration, Northern Trail Outfitters' original Bulk API (not Bulk API 2.0) jobs report some records failed with lock errors, and a few batches were processed several times before finishing. The integration lead must write the retry procedure and asks the data architect how Bulk API handles record locks. Which two statements should the architect give? Choose 2 answers.
A. When a batch is put back in the queue because of lock problems, records already marked as failed aren't retried and must be resubmitted in a separate batch
B. A batch can be requeued and reprocessed up to 10 times before it is permanently marked as failed, and some of its records can still have succeeded
C. Bulk API returns an error as soon as a record is locked, so every lock failure in the results happened on the first attempt at that record
D. A batch that is permanently marked as failed is rolled back completely, so the whole batch can be resubmitted without creating duplicates
E. Bulk API switches a job to serial mode automatically after the first lock error, so later batches in the job can't fail because of locks

Accepted Answer

Correct answer: A. When a batch is put back in the queue because of lock problems, records already marked as failed aren't retried and must be resubmitted in a separate batch | B. A batch can be requeued and reprocessed up to 10 times before it is permanently marked as failed, and some of its records can still have succeeded
Salesforce's Bulk API guidelines say that if more than 100 records in a batch can't get locks, the rest of the batch goes back in the queue, and when it is processed again the records already marked as failed aren't retried, so they must be submitted again in a separate batch. A batch can be requeued and reprocessed up to 10 times before it is permanently marked as failed, and even then some records could have completed successfully. Bulk API waits a few seconds for a lock to be released before failing a record, so it doesn't fail on first contact. Failed batches aren't rolled back, which is why blind resubmission can duplicate records. Concurrency mode is set on the job; Bulk API doesn't switch it automatically.

Free Salesforce Platform Data Architect
Practice Questions

What is the Platform Data Architect exam like?

Questions by exam section

Data Modeling/Database Design

Question 1 · Medium

Question 2 · Medium · Choose 2

Question 3 · Medium · Choose 2

Question 4 · Easier

Master Data Management

Question 5 · Medium

Salesforce Data Management

Question 6 · Harder

Question 7 · Medium

Question 8 · Easier

Question 9 · Harder

Data Governance

Question 10 · Medium

Question 11 · Medium · Choose 3

Large Data Volume Considerations

Question 12 · Harder

Question 13 · Easier · Choose 2

Question 14 · Medium

Question 15 · Easier

Data Migration

Question 16 · Harder

Question 17 · Harder · Choose 2

Question 18 · Harder · Choose 2

Where to go after these 18 questions

Get Platform Data Architect exam updates by email

Ready for a full Platform Data Architect practice exam?

Free Salesforce Platform Data ArchitectPractice Questions