Choosing a fabric data store

Choosing a Fabric Data Store 

OneLake, Warehouse, Lakehouse, Datamart, and Eventhouse are the main data storage solutions in Microsoft Fabric. In this article, we will take a deep dive into defining each of these storage solutions, examining their unique features. This guide will help you make the right decision on where to store your data while moving to Microsoft Fabric. 

 OneLake 

OneLake is a single, unified, logical data lake for your whole organization. It’s the OneDrive for data. OneLake comes automatically with every Microsoft Fabric tenant and is designed to be the single place for all your analytics data. 

Warehouse 

A lake-centric data warehouse built for any skill level (Citizen developer, DBA to Data Engineer) allowing a centralized repository that can handle very large volumes of data as a SaaS environment. 

Data is ingested into warehouse though pipelines dataflows, Cross database Querying or COPY INTO commands. Once ingested, it can be analyzed by different groups though sharing or databases querying.  

Lakehouse 

A lakehouse is a data architecture platform that allows storage, management and analysis of both structured and unstructured data in Microsoft Fabric. Lakehouse, like warehouse, has the capacity to handle a very large volume of data.  

Data is ingested into lakehouse via File upload from local computer (automatically determines the file structure and generates metadata for it), Pipelines, dataflows and Spark libraries in notebook. Lakehouse also has the capability to set up shortcuts that allows users to reference data without having to copy or create a separate copy.   

Lakehouse allows data streaming which is a continuous data ingesting as the data becomes available.  

Datamarts 

Datamarts are self-service analytics solution that enables users to explore data loaded in a managed database in a web-based environment. 

Being a web-based solution, Datamarts are simple and require no-code experience to ingest data. Data is ingested from different sources, can be cleansed using an ETL process in Power Query then thereafter loaded into a fully managed Azure SQL database.  

Eventhouse/KQL 

Eventhouse is a solution designed to handle and analyze large volumes of data particularly when near real-time analytics and exploratory is a requirement. They are the preferable engine for handling unstructured and semi structured data like free text analysis. It’s also designed as a workspace of databases allowing it to be sharable across projects and managed all at once.  

Data is ingested into KQL database as the data store.  

Now that we have some understanding of the storage elements in Fabric, here is the side-by-side comparison of these data stores