Press "Enter" to skip to content

Curated SQL Posts

Contrasting LTAP and HTAP

Published 2026-09-23 by Kevin Feasel

Paul Andrew notes another convergence of OLTP and OLAP:

At the Data and AI Summit in June this year, Databricks introduced Lake Transactional/Analytical Processing, or LTAP. My first reaction, I’ll admit, was cynicism, just for a change. Every software vendor must invent new names for old things these days. And for those of us who have worked in data architecture long enough to have the grey hair to prove it, like me, this felt very familiar. Hybrid Transactional and Analytical Processing, or HTAP, a term Gartner coined back in 2014, has long advertised bringing operational transactions and analytics “closer together”. In my experience it never really became a strong implementation pattern in data platform deliveries.

Read on to learn how LTAP and HTAP differ, where Databricks and other competitors (like Microsoft) are going in this space, and how much they have yet to do.

Leave a Comment

Vector Comparison in SQL Server 2025

Published 2026-09-23 by Kevin Feasel

Chad Callihan demonstrates a vector comparison:

Vectors consist of an ordered list of numbers that represent characteristics or meaning captured from data. That data could be words, images, or other files. An AI embedding model is used to analyze the data and translate it into a long list of numbers that becomes that data’s vector embedding. The closer the vector embeddings are to each other, the more closely related the meanings of the data are together.

While vectors are more commonly hundreds or thousands of numbers, we’re going to simplify it and make up three numbers for our example.

Click through for that comparison.

Leave a Comment

Chain-Ladder Reserving Calculations in Python

Published 2026-09-22 by Kevin Feasel

Christian Lorentzen digs into loss reserving:

Ask a reserving actuary how they run a Chain-Ladder and you’ll usually hear “Excel” or the name of a pricey specialized tool. It turns out a modern dataframe library handles it just as well — in a few lines, for hundreds of companies at once.

We use the CAS loss reserving data, specifically the other liability line of business (LoB): 233 US insurers (“GRNAME”), 10 accident years (1998–2007), paid and incurred losses at every development lag (1-10). We treat 2007 as our reporting year, i.e. we simulate a year-end closing.

Click through for a demonstration and a comparison against R’s ChainLadder package.

Leave a Comment

Dealing with Bias in Data Stories

Published 2026-09-22 by Kevin Feasel

Cole Nussbaumer Knaflic provides some advice:

Yesterday, a participant in one of our workshops asked a question that I’ve been thinking about ever since. She works with researchers and statisticians, and the idea of telling a story with data can make some of them uncomfortable. Doesn’t telling a story mean choosing a perspective? And doesn’t choosing a perspective introduce bias? Wouldn’t it be more objective to simply show the data and let people draw their own conclusions?

Read on for Cole’s answer. Cole differentiates between “bias” and “perspective,” though I wouldn’t even go that far. Instead, I’d simply make it clear what your biases are. At its core, a bias is a preference for or against something. We normally think of bias as a negative, but that’s not strictly true. I do think it’s important to state where your biases come in, as they can undermine a point or attempt to drive people to a specific conclusion. But bias itself can extend to “I am showing you this information because I believe it will help you in making your decision,” even if that decision isn’t one you necessarily prefer.

Leave a Comment

Shortcuts to Iceberg Tables in Microsoft Fabric

Published 2026-09-22 by Kevin Feasel

Teo Lachev makes a connection:

Snowflake is an increasingly popular option for data warehousing. One nice Snowflake feature is the ability to configure a table to save its data in Iceberg file format in the Snowflake own managed storage or external storage, including S3, GCS, ADLS Gen 2, and OneLake. As I wrote in the post “Give Me Your Data!”, Fabric supports shortcuts to Iceberg tables, in which case the data is not copied but exposed as a Parquet Delta table in a Fabric lakehouse. This could be useful when building Fabric-centric solutions on top of Snowflake without moving the data.

Iceberg isn’t just a Snowflake thing, though this is certainly helpful if you do have both Snowflake and Fabric.

Leave a Comment

Migrating a Power BI Report between Tenants

Published 2026-09-22 by Kevin Feasel

Reitse Eskens makes a move:

Before I get started, yes, I’m going to write about Power BI. If you’re not familiar with my work, the long and short of it is that I tend to keep a healthy distance from anything involving semantic models and reports. Why? In one word, pie charts. Sorry if that triggered you.

But every now and again, I need to do work adjacent to this; specifically, when onboarding new clients or helping them migrate things, I have to touch the Power BI stuff too. But rest assured, a lot needs to happen before I’ll actually start creating reports. Many people are more qualified than me to do that.

Click through to see how you can do this through the Fabric/Power BI portal. And if you want to see this made easier, vote up Reitse’s suggestion.

Leave a Comment

Setting Fabric Connection Properties with the ODBC Driver

Published 2026-09-22 by Kevin Feasel

Ryan Adams digs into configuration settings:

Some SQL Server ingestion scenarios depend on connection-string properties that the built-in SQL Server connector in Data Factory for Microsoft Fabric doesn’t expose. If you need to route reads to an Availability Group secondary, connect reliably across a multi-subnet listener, or read columns protected by Always Encrypted, you’ve probably hit that wall. This post shows a supported way around it: use the ODBC connector together with an ODBC data source name (DSN) configured on an on-premises data gateway (OPDG). The example in this post uses Always Encrypted in a Copy job.

Click through to see some of the things you can set. These are standard SQL Server properties, so they can also work with normal SQL Server installations that are making use of Availability Groups and Always Encrypted.

Leave a Comment

Configuring Row-Level Security with OneLake Security

Published 2026-09-22 by Kevin Feasel

Reza Rad secures some data:

If you’ve set up row level security in Power BI before, you know the usual drill: open the semantic model, add a role, write a DAX filter. But once your data lives inside a Microsoft Fabric OneLake structure, there’s a better way to do this. You can implement it directly in OneLake instead, so that every object built on top of that data, your lakehouse, your SQL analytics endpoint, your semantic model, and your report, all follow the same security setup automatically. Define it once, upstream, and everything downstream inherits it. This is called OneLake security, and in this post I’ll walk through exactly how to set up row level security (RLS) this way, from the lakehouse all the way to your Power BI report.

Read on for a video and summary with timestamps.

Leave a Comment

Broker Tasks on Query Store Cleanup

Published 2026-09-21 by Kevin Feasel

Erik Darling does some digging:

While looking at a client’s server, I noticed there was, at any given time, 2-3 background sessions with the command column saying BRKR TASK, which I found quite odd.

Only the Microsoft-shipped broker queues were present, and none were activated. Nothing uses Service Broker, Mirroring, or Log Shipping.

Read on for Erik’s findings. If you need a Misfits song to listen to, we are close enough to the end of October for Halloween to work.

Leave a Comment