When dealing with Microsoft Azure Cosmos DB, understanding when to use a point operation versus a query operation is crucial. While both operations have their specific use cases, yet knowing which one to employ when can greatly optimize your results or improve your application’s performance.
1. Point Operations
Point Operations in Azure Cosmos DB are specific read and write operations on a single item identified by its key. In simpler terms, you are pinpointing exactly what item you want to interact with based on its unique identifier (id and partition key). Point operations include:
- Create
- Read
- Replace
- Upsert
- Delete
These operations significantly benefit from lower latency and pricing due to their specific nature. They are the preferred path for CRUD operations on individual items.
For instance, using C# SDK v3 to read an item in your Cosmos DB corresponds to the following code:
ItemResponse
Item item = response.Resource;
Where “itemId” and “partitionKeyValue” are id and partition key of the document, respectively.
2. Query Operations
On the other hand, query operations in Azure Cosmos DB allow you to exploit the power of SQL against your NoSQL database. Below are some types of Query operations,
- Queries that return all the properties
- Range queries
- Join operations
Unlike point operations that target a single unit, query operations potentially engage multiple units, resulting in higher request unit charges. Therefore, they are recommended when you need to retrieve more than one item,
Below is an example of a query operation that retrieves all the families with at least one child:
string queryStr = "SELECT * FROM c WHERE " +
"ARRAY_LENGTH(c.children) > 0";
FeedIterator
In this code, we are using the GetItemQueryIterator method to pass our SQL query and retrieve families with at least one child.
Comparison between Point Operations and Query Operations
| Operation | Latency | RUs (Request Units) | Scope | Use cases |
|---|---|---|---|---|
| Point Operations | Low | Low | Single item | CRUD operations on individual items |
| Query Operations | Higher than point operations | Depends on the query type and complexity | Multiple items | Retrieve multiple items |
Both point and query operations have unique benefits and potential drawbacks. The deciding factor on which to use comes down to the specific requirement and application use-case. Employing these understanding of these operations will enable you to make an informed decision, reduce unnecessary costs and drastically improve the performance and efficiency of your Microsoft Azure Cosmos DB-based applications.
Remember, choosing between a point operation and a query operation is not about which is better — it’s about which is right for your specific use-case. In the DP-420 Designing and Implementing Native Applications Using Microsoft Azure Cosmos DB exam, you will find a strong emphasis on these concepts, underlining their importance in real-world applications.
Practice Test
True or False: Point operations are ideal when you need to fetch a large amount of data.
• True
• False
Answer: False.
Explanation: For fetching a large amount of data, query operations are preferable. Point operations are best for fetching specific data.
True or False: The cost of a point read operation is always fixed.
• True
• False
Answer: True.
Explanation: The cost of a point read operation is fixed at 1 request unit, making it very predictable.
True or False: Query operations are always less expensive than point operations.
• True
• False
Answer: False.
Explanation: The cost of a query operation varies based on a number of factors, including the query complexity, dataset size, and throughput.
Which of the following operations should be used when you need to retrieve a list of all employees in a certain department?
• a) Point operation.
• b) Query operation.
Answer: b) Query operation.
Explanation: Since this requires the retrieval of multiple items that satisfy a specific condition (being in a certain department), a query operation is appropriate.
When should one use a point operation?
• a) When you need to fetch multiple items based on a complex condition.
• b) When you are unsure of the partition key.
• c) When you need to fetch a specific item with a known partition key and id.
• d) When you need to fetch a large amount of data.
Answer: c) When you need to fetch a specific item with a known partition key and id.
Explanation: Point operations are ideal when you need to fetch a specific item and you know the partition key and id.
True or False: Query operations always return a single item with a known partition key and id.
• True
• False
Answer: False.
Explanation: Query operations return multiple items that satisfy a condition and are not limited to a single item with a known partition key and id.
When should a query operation be used?
• a) When data size is small.
• b) When retrieving a single item by its id.
• c) When fetching multiple items based on a complex condition.
• d) When partition key is unknown.
Answer: c) When fetching multiple items based on a complex condition.
Explanation: Query operations should be used for fetching multiple items that satisfy a specified query or condition.
True or False: Query Operations in Azure Cosmos DB can take advantage of indexing to improve performance.
• True
• False
Answer: True.
Explanation: Azure Cosmos DB utilizes automatic indexing which can significantly improve query performance.
True or False: Point operations offer single document transaction semantics.
• True
• False
Answer: True.
Explanation: Point operations offer single document transaction semantics which means they are atomic within the scope of a single partition key.
Which one is more suited for complex searches or filters?
• a) Point operation.
• b) Query operation.
Answer: b) Query operation.
Explanation: Query operations are more suited for complex searches or filters since they can operate over a larger set of data and apply more complex search conditions.
True or False: Point operations in Azure Cosmos DB support Server-side JavaScript (JS) stored procedures.
• True
• False
Answer: True.
Explanation: Point operations fully supports Server-side JavaScript (JS) stored procedures, triggers, and user-defined functions (UDFs).
Choose the type of operation with higher throughput but variable costs.
• a) Point operation.
• b) Query operation.
Answer: b) Query operation.
Explanation: While query operations have higher throughput, they come with variable costs depending on complexity or the amount of data they access.
True or False: Point operations are idempotent operations.
• True
• False
Answer: True.
Explanation: Point operations are idempotent, meaning making the same multiple requests results in the same single outcome.
Which of the following operation types can efficiently handle batch operations?
• a) Point operation.
• b) Query operation.
Answer: a) Point operation.
Explanation: Point operations can efficiently handle batch operations or transactions on multiple items within the same partition.
In Azure Cosmos DB, which operation adheres to the ACID properties for all operations within a specific partition key?
• a) Point operation.
• b) Query operation.
Answer: a) Point operation.
Explanation: Point operations fully conform to ACID principles within the scope of a single partition key.
Interview Questions
What is a point operation in Microsoft Azure Cosmos DB?
A point operation in Microsoft Azure Cosmos DB refers to operations that interact with a specific item. These operations include creating, reading, replacing, or deleting an item in a collection.
What is a query operation in Microsoft Azure Cosmos DB?
A query operation in Microsoft Azure Cosmos DB involves executing a SQL query against a collection of items. These operations can filter, project, sort, join, and aggregate data across multiple items.
When is it more suitable to use a point operation?
Point operations are best used when you already know the item’s id and/or the partition key. These operations are also extremely efficient for both performance and cost, as they leverage the partition key.
When should we consider using a query operation?
Query operations should be considered when you need to interact with multiple items that have different ids or partition keys. Queries can filter, project, sort, join, and aggregate data across multiple items, enabling complex data manipulations.
What is the benefit of using a point operation in Cosmos DB?
Point operations are very efficient as they consume the least Request Units (RUs) among all operation types. They involve direct operations on an item using its id and partition key, which leads to optimized performance and less resource consumption.
How does Cosmos DB charge for a point write operation?
Cosmos DB charges for a point write operation based on the size of the item. For example, a 1 KB document will consume 5.5 Request Units (RUs).
How does Cosmos DB charge for a query operation?
Cosmos DB charges for query operations based on the number of items returned, the amount of data processed, and the complexity of the query.
Can we perform bulk operations using point operations in Cosmos DB?
No, point operations in Cosmos DB only allow operations on a single item at a time. For bulk operations, it is better to use bulk API support.
What is the role of an indexing policy in query operations in Cosmos DB?
An indexing policy in Cosmos DB controls the aspects of indexing for a container. It determines how items should be indexed and influences the performance and cost of queries.
What is the role of partition key in point operations in Cosmos DB?
The partition key in point operations in Cosmos DB helps distribute data evenly and ensures that the operation is efficient. It allows direct access to an item, thereby reducing latency.
Can a query operation be executed without a partition key in Cosmos DB?
Yes, a query operation can be executed without specifying a partition key. However, this might lead to cross-partition queries which can consume more resources and lead to higher latency.
Can you use transactional batch for point operations in Cosmos DB?
Yes, Azure Cosmos DB provides transactional batch support for point operations. It allows you to execute several point write operations on items with the same partition key, within a single transaction.
In Azure Cosmos DB, does changing a point operation into a query operation change the cost?
Yes, changing a point operation into a query operation can change the cost. Generally, query operations are more costly than point operations due to the increased complexity and resources required to process the operation.
What performance implications can occur if query operations are not correctly optimized in Cosmos DB?
If query operations are not correctly optimized, it could lead to resource throttling, increased latency, and higher cost due to the consumption of more Request Units.
How can you optimize a query operation in Cosmos DB?
Query operations can be optimized in Cosmos DB by using indexing policies, choosing the appropriate consistency level, using pagination, and avoiding cross-partition queries wherever possible.
extutorials.com