C100DEV Data Modeling Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Embed the `medications` array as is, and create a multikey index on `medications.name` and `medications.dosage`.

  Why it's wrong here

  While a multikey index can support queries on array elements, embedding hundreds of medication subdocuments per patient can lead to large documents and potential performance issues. Each index entry is created for each array element, which can bloat the index. Additionally, updating or querying specific medications within a large array can be inefficient. This approach does not scale well for hundreds of entries.

• ✓

  Move medications to a separate `medications` collection with documents referencing the patient ID, and create a compound index on `{ name: 1, dosage: 1, patientId: 1 }`.

  Why this is correct

  Separating medications into their own collection avoids unbounded array growth in patient documents and allows for efficient querying with a compound index on `name`, `dosage`, and `patientId`. This design supports queries for specific medication and dosage across patients, and it scales well as the number of medications grows. It also simplifies updates and deletions of individual medication records.

• ✗

  Use the Subset Pattern to store only the most recent 10 medications in the patient document and archive older medications in a separate collection.

  Why it's wrong here

  The Subset Pattern is useful when only a subset of data is frequently accessed, such as recent orders. However, in a healthcare context, all medications may be relevant for queries, not just the most recent ones. Archiving older medications could lead to incomplete query results. This pattern does not address the need to query any medication by name and dosage efficiently.

• ✗

  Use the Bucket Pattern to group medications into buckets of 50 per patient document, and index the `name` and `dosage` fields.

  Why it's wrong here

  The Bucket Pattern is typically used for time-series data, not for grouping related subdocuments like medications. While it reduces the number of documents, it complicates queries that need to find a specific medication across buckets. The application would need to query multiple buckets, and the index would still need to cover all elements. This adds complexity without clear benefits.