Salesforce Certified Platform Data Architect Free Sample Questions

A Master-Detail relationship forces the child record to inherit the sharing settings of the parent, which violates the security requirement that access be strictly controlled independent of the parent Account. Standard Roll-up Summary fields are only available on Master-Detail relationships. Therefore, to satisfy the security requirement (Lookup relationship for independent security) and the reporting requirement (Roll-up), the architect must use a Lookup relationship and implement the roll-up logic via a Trigger or Declarative Flow.

Skinny tables are a powerful performance optimization tool for custom objects with large data volumes (LDV). They improve performance by avoiding joins (combining fields from the main storage table and the custom field table) and by excluding soft-deleted records. Since the object has millions of records and standard indexing is insufficient, a skinny table is the next logical on-platform optimization step.

In parallel mode, several batches insert child records of the same parent Account at the same time, and saving a child record locks its parent, so batches wait for each other and fail with UNABLE_TO_LOCK_ROW. The Bulk API guide's advice is to organize the data so that records referencing the same parent are in a single batch: sorting the CSV by the parent Account ID groups each Account's orders together, so parallel batches touch different parents and throughput stays high. Serial mode also avoids the locks, but it processes one batch at a time; the guide says to avoid it unless you can't reorganize your batches. The SOAP API is slower for 5 million records and doesn't remove the parent locks, and deactivating triggers speeds processing but doesn't stop the parent lock. Sources: Bulk API Developer Guide (Spring '23), General Guidelines for Data Loads; Designing Record Access for Enterprise Scale, Record-Level Locking.

Big objects are built for this pattern: the Big Objects Implementation Guide lists auditing and tracking, and historical archive for analysis or compliance purposes, as custom big object use cases, with consistent performance at hundreds of millions or billions of records. Records are queried with SOQL through the big object's index, there is no standard edit UI (only custom Lightning or Visualforce components), and access is controlled by object and field permissions, so standard users can get read access without any right to change or delete the logs. A custom object with read-only page layouts can still be edited through the API, and about 5 billion records would overwhelm data storage; Task records have the same storage problem. An external object only displays data that lives in another system, so the logs would have to be stored and secured outside Salesforce. Source: Big Objects Implementation Guide (Spring '23), Custom Big Object Use Cases and Considerations.

The scenario describes a Many-to-Many relationship (One Shipment -> Many Containers; One Container -> Many Shipments). In Salesforce, this is modeled using a Junction Object. The requirement to track the 'assignment date' further supports this, as the date is an attribute of the relationship itself, which would be stored on the junction object.

Salesforce Connect allows you to view and search data stored outside Salesforce (Data Virtualization) via External Objects. This meets the requirement of not storing data in Salesforce while enabling real-time access and reporting capabilities (limitations apply but basic reporting is supported).