Azure DP-900: Microsoft Azure Data Fundamentals | Notes and Highlights

Just passed Azure DP-900 exam!

Here is what and how I did it:

• Get all modules done, read and exercises on Microsoft Learn for DP-900, link.
• Make sure you have solid understanding of all topics covered in the sample questions, link. Not just the question and answers, but also explore all the url given in those response boxes.
• Also, highly recommend to read this article: https://learn.microsoft.com/en-us/azure/architecture/guide/technology-choices/data-store-overview

And here are all notes I’ve taken during my study process:

Document Data Store

• All data should be stored and retrieved from a single-data-structure
• In a single document, use to store user profile data, like shipping addresses, user preferences
• Each user can have different number of addresses and also different number of preferences
• Document database consists of entities that have their related data stored in a single document
• Documents supports flexible schema
• Usually JSON format documents
• Document data store is a type of a non-relational data store

Azure Storage

• Enable hierarchical namespace to implement folder and directory level security in Azure Storage.
• Hierarchical namespace allows you to organize blob containers in folders and directories
• Hierarchical namespace allows to define PSIX compatible permissions to container
• Hierarchical namespace allows to define RBAC to container

Azure SQL Database

• Azure SQL Database has lower cost compared against Azure SQL Managed Instance
• Azure SQL Database has minimal administration compared against Azure SQL Managed Instance
• Azure SQL Database has higher availability ( 99.995% ) compared against Azure SQL Managed Instance ( 99.99% )
• Azure SQL Database is a relational database for modern cloud apps
• Azure SQL Database is PAAS
• Azure SQL Database eliminate the need to manage the SQL Server configuration
• Azure SQL Database is not as compatible to on-premises as Azure SQL Managed Instance is
• Azure SQL Database does not allow you to choose the SQL Server engine version
• Azure SQL Database uses the lasted stable version of the SQL Server engine
• Azure SQL Database includes a fully managed backup service
• Azure SQL Database automatically applies security patches on underlying operational system

Graph Database

• NODE is an instance of data entities, like people on companies
• EDGE are relationships between instances of data entities
• PROPERTY are attributes of data entities, like person email, address
• DIRECTIONS are property of an Edge. Edges can be directed or undirected
• Graph database store semi-structured data

Azure HDInsight

• Azure HDInsight is a big data processing service
• Azure HDInsight is used to provision and manage a cluster of open-source analytics solution
  • like Apache spark, Hadoop, Kafka
• Azure HDInsight is compatible with Apache Spark

Azure Data Explorer

• Uses KQL language

Azure Table Storage

• Azure Table Storage is composed by row key + partition key
• Azure Table Storage number of columns may vary according to each row
• Azure Table Storage is a NoSQL data store, uses unstructured data
• Azure Table Storage is a key/value data store
• Azure Table Storage must have a unique key ( row key + partition key combination )
• Azure Table Storage number of column per row can vary
• Azure Table Storage partition key: identify partition containing in the row
• Azure Table Storage row key: is unique to each row in the same partition

Key / Value Data Store

• Key / Value data store can receive real-time events from multiple sources which may vary in schema
• Key / Value data store works like a lookup
• Key / Value data store find rows by using unique key
• Key / Value data store is essentially a large hash-table
• Key / Value data store is optimized for fast data writing
• Key / Value data store each data row is referred by a single key-value

Spark Structured Streaming

• Spark Structured Streaming can be used with: Azure Synapse Analytics, Azure DAtabricks, Azure HDInsight
• Spark Structured Streaming is open-source library for apache-spark based services
• Spark Structured Streaming is a near real-time processing engine

Azure Cosmos DB

• Azure Cosmos DB allows simpler queries to retrieve data than a relational DB
• Azure Cosmos DB supports app with global user base
• Azure Cosmos DB is a NoSQL database
• Azure Cosmos DB supports storing data in denormalized way
• Azure Cosmos DB suppoarts multi-region writes
• Azure Cosmos DB general resource hierarchy is: 1 azure cosmos account > 2 database > 3 container > 4 item