Press "Enter" to skip to content

Day: April 8, 2026

Accelerated Database Recovery in tempdb for SQL Server 2025

Published 2026-04-08 by Kevin Feasel

Rebecca Lewis looks into a feature:

Two weeks ago I covered the Resource Governor changes in SQL Server 2025 — specifically, capping how much tempdb data space a workload group can consume. That was the data-file side. For the log side, SQL Server 2025 now lets you enable Accelerated Database Recovery (ADR) on tempdb. Enable it and cancelled queries stop grinding, the tempdb log stops bloating, and recovery gets faster. Sounds like an easy yes — but you’ve got to read the fine print.

Click through for that fine print.

Leave a Comment

A Gotcha with pg_column_size()

Published 2026-04-08 by Kevin Feasel

Laetitia Avrot wants to know how large a column is:

Thanks to my colleague Ozair, who sent me a JIRA ticket saying “I need to drop that huge column, what are the consequences?” My first question was: how huge? That’s when the rabbit hole opened.

It looks simple. It is simple. Just use the administrative function pg_column_size(). Until you have toasted attributes. Then it gets interesting.

Read on for a demonstration.

Leave a Comment

Shortcut Transformations now GA in Microsoft Fabric

Published 2026-04-08 by Kevin Feasel

Pernal Shah transforms some data:

Organizations today manage data across multiple storage systems, often in formats like CSV, Parquet, and JSON. While this data is readily available, turning it into analytics-ready tables typically requires building and maintaining complex ETL pipelines.

Shortcut transformations remove that complexity.

With Shortcut transformations, you can convert structured files referenced through OneLake shortcuts into Delta tables without building pipelines or writing code.

This currently works for CSV, Parquet, and JSON data and does cut out a very common step for raw-layer transformation.

Leave a Comment

When Multi-Column Statistics (Mostly) Aren’t

Published 2026-04-08 by Kevin Feasel

Brent Ozar digs into multi-column statistics:

The short answer: in the real world, only the first column works. When SQL Server needs data about the second column, it builds its own stats on that column instead (assuming they don’t already exist), and uses those two statistics together – but they’re not really correlated.

For the longer answer, let’s take a large version of the Stack Overflow database, create a two-column index on the Users table, and then view the resulting statistics:

Click through for a dive into row estimation when you have multiple columns in a WHERE clause and how statistics come into play.

Leave a Comment

TOP(1) with Ties

Published 2026-04-08 by Kevin Feasel

Andy Brownsword can’t stop at one:

Having TOP (1) return multiple rows feels wrong… but that’s what WITH TIES can do.

For a long time I used patterns like this to get the first record in a group:

Andy goes on to explain how WITH TIES works in T-SQL, shows an alternative to using a common table expression + window function to narrow down to the first logical group, and digs into when you might not want to use that alternative.

Leave a Comment