Press "Enter" to skip to content

Curated SQL Posts

Random Page Cost and PostgreSQL Query Plans

Tomas Vondra takes us through a setting:

Last week I posted about how we often don’t pick the optimal plan. I got asked about difficulties when trying to reproduce my results, so I’ll address that first (I forgot to mention a couple details). I also got questions about how to best spot this issue, and ways to mitigate this. I’ll discuss that too, although I don’t have any great solutions, but I’ll briefly discuss a couple possible planner/executor improvements that might allow handling this better.

Tomas’s points around the random_page_cost setting sound a lot like the cost threshold for parallelism setting in SQL Server in inverse: a setting whose default makes sense in a world of spinning disks at 7200 RPM, but not really in a solid state world.

Leave a Comment

The Importance of Power BI Performance Load Testing

Gilbert Quevauvilliers runs some tests:

It is becoming increasingly important to understand how the Power BI reports/Semantic Model that are being used in your organization are performing.

When using Fabric Capacities this can potentially be of critical importance, because a single report that is not well designed could cripple or bring down your capacity.

By completing Power BI Performance load testing before it goes into a production environment allows for scalable, dependable, repeatable testing to take place in lower environments.

Read on to see what this entails and the tool Gilbert will use throughout this series.

Leave a Comment

Writing Back to a Fabric Data Warehouse via UDF

Jon Vöge continues a series on write-back options into Microsoft Fabric:

In that article, we took advantage of some of the built-in sample code from the User Data Function editor, as well as some great code examples from Sujata: Example User data functions for Translytical task flows · GitHub

The problem? All of these samples use SQL Databases in Fabric as the backend item.

Jon switches this from a SQL database into a Fabric Data Warehouse, and notes some of the challenges along the way.

Leave a Comment

k Nearest Neighbors Search in Elasticsearch

Govind Singh Rawat looks for nearby documents:

Businesses are increasingly relying on intelligent search capabilities to enhance customer experience, automate insights, and unlock the potential of unstructured information. Elasticsearch, a leading distributed search and analytics engine, is at the heart of many such systems. One of its powerful and lesser-known capabilities is support for k-nearest neighbors (k-NN) search, a method particularly useful for similarity-based retrieval in domains such as semantic search, recommendation engines, and image recognition.

This article delves into what Elasticsearch and k-NN search are, how the two are integrated, and how to configure and optimize k-NN in Elasticsearch for real-world applications.

Click through for a high-level primer on the topic, as well as a few links to additional resources.

Leave a Comment

Event Streaming in Microsoft Fabric

Rayis Imayev streams some data:

In my post last week (https://datanrg.blogspot.com/2025/06/salesforce-cdc-data-integration.html), I talked about Salesforce Change Data Capture (CDC) event data streaming, where the initial event destination was file storage in Azure. But what if we anticipate a higher volume of incoming Salesforce source data or the addition of a new data feed? This could create the need for an alternative method of managing incoming events.

Read on to learn more.

Leave a Comment

Troubleshooting Network-Related or Instance-Specific Error

Aaron Bertrand has started a new series:

This is the first in a series of articles meant to provide practical solutions to common issues. In this post, we’ll talk about one of the most pervasive error messages out there:

A network-related or instance-specific error occurred while establishing a connection to SQL Server. Verify that the instance name is correct and that SQL Server is configured to allow remote connections.

Read on to see what a variety of potential solutions to this problem. I was going to joke “It’s always DNS” but Aaron actually has a section on DNS in there.

Leave a Comment

Bioconductor in the Wake of ggplot2 4.0.0

Maria Doyle lays it out:

A major update to ggplot2 (version 4.0.0) is expected around mid-to-late July 2025. It brings a significant internal change, replacing most of the S3 backend with the newer S7 object system. While this improves long-term maintainability and extensibility, it may break Bioconductor packages that depend on ggplot2, especially those that customise how plots are built or styled. Packages that use ggplot2 for typical plotting tasks, such as creating plots with ggplot() and geom_*(), are unlikely to be affected.

Click through for notes, tips on what to do, and whether the code you’re using will break with ggplot2 4.0.0. H/T R-Bloggers.

Leave a Comment

Trying out Fabric Unified Admin Monitoring

Reitse Eskens tries out a tool for monitoring Microsoft Fabric installations:

Let me set the scene quickly for you. You’re working for an organisation where Fabric is in the process of being adopted or it’s already fully in use. Regardless of the number of capacities, workspaces, etc, you’re interested in what’s going on in your Fabric environment. You have questions like “Who is using the reports?”, “Who is changing settings in the Admin panel?” and “How is my capacity being used?”.

Read on for a single tool that can solve these sorts of questions.

Leave a Comment

Drop Shadows and Power BI

Elena Drakulevska has some thoughts on drop shadows:

I get why people add them. Shadows might feel like a design upgrade. A quick way to make your visuals pop or feel more “finished.”

But here’s the thing: just like rounded corners, drop shadows are easy to overdo—and they’re not actually helping. Not with clarity. Not with accessibility. Definitely not with UX.

Click through for Elena’s full thoughts. I’m generally against drop shadows. They draw visual attention without providing the report viewer any value. That’s chartjunk.

Leave a Comment