Bitzu

AZ-305 ยท Data integration and analysis

Choosing an analytics platform

Microsoft Fabric vs Azure Synapse Analytics vs Azure Databricks for data warehousing, big data and ML; dedicated vs serverless SQL; Power BI for reporting.

Key points

  • Microsoft Fabric: SaaS (Data Factory, Spark, Data Warehouse, Power BI) over OneLake: one per tenant, ADLS Gen2, Delta Lake, shortcuts to S3 and others; nothing to provision.
  • Synapse dedicated SQL pool: MPP, sized in DWUs, 60 distributions (hash, round-robin, replicated), can pause.
  • Synapse serverless SQL pool: T-SQL over lake files, scales automatically.
  • Azure Databricks: managed Spark, Delta Lake, ML; Standard and Premium SKUs.

Source: https://learn.microsoft.com/en-us/fabric/fundamentals/microsoft-fabric-overview

Practice questions

Single choice

Which Azure Synapse Analytics option lets analysts run T-SQL directly over Parquet files in the data lake without provisioning any compute?

  1. Dedicated SQL pool
  2. Apache Spark pool
  3. Copy activity in a pipeline
  4. Serverless SQL pool
Show answer

Answer: D. Serverless SQL pool

A serverless SQL pool scales automatically per query and reads lake files in place. A dedicated SQL pool must be provisioned in DWUs, a Spark pool runs Spark rather than T-SQL, and a Copy activity moves data instead of querying it.

Select all that apply

Which statements correctly describe a Synapse dedicated SQL pool?

  1. It scales automatically per query with no capacity to provision
  2. Its compute can be paused while the stored data is kept
  3. It can only query external files and cannot ingest data
  4. Its compute size is expressed in data warehouse units
  5. Its queries are divided across 60 distributions
Show answer

Answer: B. Its compute can be paused while the stored data is kept, D. Its compute size is expressed in data warehouse units, E. Its queries are divided across 60 distributions

A dedicated SQL pool splits every query across 60 distributions, is sized in data warehouse units, and can be paused so you pay only for storage. Automatic per-query scaling with nothing to provision describes the serverless SQL pool, and a dedicated pool can ingest data, not only query external files.

True or false

Before you can store data in Microsoft Fabric, you must create an Azure storage account to host OneLake.

Show answer

Answer: False

False. OneLake comes with every Fabric tenant with no upfront provisioning; it is built on ADLS Gen2 but managed by the SaaS platform, and you do not even need an Azure account to use Fabric. Existing storage accounts can be added through shortcuts if you want them.

Practice 8 questions on this topic in Bitzu

Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.

Download on the App StoreGet it on Google Play

More AZ-305 topics in this module