Warehouse loading – tips and tricks for better performance

Proposed session for SQLBits 2026

TL; DR

We will talk about how to combine the toolset we have for building elegant data loading solutions in the warehouse for the best performance. Tips and tricks on the way will help you learn, design and be prepared to start a process of loading data in your Fabric warehouse environment.

Session Details

Microsoft Fabric offers built-in data ingestion tools that allow users to ingest data into warehouses at scale using code-free or code-rich experiences. In this session we will discover all the possible options - T-SQL features, and code free tools, and even more that some source types provide.
We will compare them according to different criteria like performance and ease of use, source volume and types, target schemas, single or cross workspaces usage. You will see them in action and understand how they work. You will get some tips on their best usage and their best combinations. But more importantly you will understand the main and the most critical factors for the design of the loading process so that you can make it fast and smooth. We will talk about how to combine the toolset we have for building elegant data loading solutions in the warehouse for the best performance. Tips and tricks on the way will help you learn, design and be prepared to start a process of loading data in your Fabric warehouse environment.

3 things you'll get out of this session

You will understand the main and the most critical factors for the design of the loading process so that you can make it fast and smooth.

Margarita Naumova's other proposed sessions for 2026

Large Tables, Big Problems: Indexing, Partitioning, and Archiving at Scale - 2026

A Deep Dive into Table Partitioning – Part 1: Mechanics and Performance - 2026

A Deep Dive into Table Partitioning – Part 2: Design, Maintenance, and Troubleshooting - 2026

SQL Server 2025 and Microsoft Fabric: Integration Patterns and Trade-offs - 2026

SQL Server 2025 Optimized Locking and ADR: Practical Internals for Real Systems - 2026

Margarita Naumova's previous sessions

Keeping historical data in tables forever – mission (im)possible!

Let’s discover some solutions for keeping historical data in the database when you receive near to impossible requirements like storing data in same tables forever and being able to edit and query them at the same time, of course keeping the response at its best. Based on a real project case the session walks you through the design process from the start to the reaching of the final solution and making the client (and developers) happy.

Modern database design (anti)patterns

We must realize that patterns which were valid 10 years ago are less likely to work now, like cursor logic, xml usage, or storing all in db v/s using NoSQL. From global industry trends to specific database patterns, this session is a combination of best practices, good and bad patterns, tips, and tricks which I give to customers in my work as a consultant.

SQL Server Table partitioning – DOs and DON’Ts

When do we get improvement in the performance, and when should we better not choosing partitioning but some other optimization techniques instead? Shall we optimize by using partitioning or yet another index?

Identify opportunities for In-Memory OLTP

The session will uncover the In-Memory OLTP architecture, the concerns about data durability and database startup and recovery as well as some important consideration on Management of in-memory objects.