Azure Data Fundamentals Flashcards & Memory Notes

Data with a fixed schema of rows and columns, such as relational tables.

Data with flexible schema tags or keys, such as JSON or XML documents.

Free-form content such as photographs, audio, or video, not organized into predefined analytical fields. Files can still have metadata; binary encoding alone does not imply unstructured data.

Transactional handles current consistent writes. Analytical favors large reads, history, and aggregates.

Platform configuration, security, backup, restore, availability, and tuning.

Ingestion, transformation, and reliable pipelines that land data for consumers.

Exploring data, building models and visuals, and answering business questions.

No. Parquet and ORC are columnar formats for analytical reads. Parquet uses a schema and can contain nested fields; inspect the contents rather than classifying every binary file as unstructured.

For objects such as images, backups, and analytical files. Blob can hold structured, semi-structured, or unstructured contents; the object access pattern drives the choice.

When you need related tables, keys, and transactional consistency rather than objects or simple key-attribute entities.

SELECT reads rows; INSERT adds, UPDATE changes, and DELETE removes rows. JOIN is a clause that combines related rows within a query, not a standalone statement. CREATE, ALTER, and DROP change database objects.

A saved query presented as a virtual table.

Suitable indexes can improve selected lookups and filters, with storage and write overhead. An index is not a guarantee that every query becomes faster.

SQL Database is PaaS for cloud apps. Managed Instance offers near-full SQL Server compatibility as PaaS.

IaaS: you manage the operating system and SQL Server.

When the app already targets PostgreSQL and you want managed PaaS.

Azure Database for MySQL Flexible Server and Azure Database for PostgreSQL Flexible Server. Azure Database for MariaDB retired on September 19, 2025; distinguish the retired service from the MariaDB engine.

Referential integrity so related rows point to valid keys.

Reusable logic packaged on the database server.

Prefer Azure SQL Managed Instance over SQL on VMs when the stem forbids OS patching.

Objects such as photographs, backups, JSON documents, and Parquet files. Blob is also a foundation for data lakes; not every object has unstructured contents.

Managed file shares with SMB or NFS semantics.

NoSQL key and attribute entities with partition and row keys.

A globally distributed multi-model database service.

MongoDB-compatible client access. Check supported commands, versions, and behavior before assuming a migration needs no changes.

CQL-style wide-column workloads against Cosmos DB.

Graph traversals when relationships are first-class.

Table or key-attribute patterns with Cosmos distribution features.

Blob for objects, Files for shares, Table for key-attribute entities.

When you need multi-table relational transactions and foreign keys.

Landing data from sources into stores for further processing.

A unified Microsoft analytics platform spanning lakehouse, warehouse, pipelines, real-time, and BI.

An Apache Spark-based analytics and AI workspace for large-scale data work.

Batch processes bounded jobs. Streaming processes continuous events with low latency.

Event Hubs ingests and buffers events. Azure Stream Analytics processes streams, for example by calculating time-window aggregates. Fabric also provides real-time analytics capabilities.

A report can contain multiple interactive pages. A dashboard is a single canvas in the Power BI service and can combine tiles from different reports. Both rely on underlying data and semantic models.

Facts record events at a defined grain; dimensions provide descriptive grouping and filtering attributes. Measures calculate results in the current filter context. Relationships connect the tables.

To show trends over time.

To compare values across categories.

When there are many categories; too many pie slices become unreadable.