Episode 310 – October 2025 Feature Summaries

Welcome to what Jason and John are calling a “Back to the Future Style podcast” – where the hosts time travel back to October to cover feature releases they missed while life got in the way. Recorded just after watching all the Microsoft Ignite announcements (more on those in upcoming episodes), the guys tackle both Power BI and Microsoft Fabric October 2025 updates in preparation for the big news drops.

As Jason noted at the top of the show: “October is a little lighter because they were getting ready for the buildup for Ignite.” But don’t let that fool you – there’s still plenty of substantial updates to discuss, from Copilot going GA to performance enhancements in Fabric’s data engineering stack.

Life Updates: Marching Bands and Emergency Rooms

Before diving into the technical updates, the hosts shared some personal context for why they’ve been MIA. Jason’s son’s marching band season wrapped up with impressive results – sixth in state and sixth at Grand Nationals. Meanwhile, John managed to add a post-midnight emergency room visit to his “conference bingo card,” which as the community friends noted, will become “another Epic Jason and John conference story.”

But the real news? They’re heading to Dublin after American Thanksgiving and have Steph Bruno lined up as a guest to discuss the Ignite announcements. Jason teased some controversy: “There’s been some visceral reactions to the keynote…some real visceral reactions. There’s some good blog posts out there about people’s viewpoints on AI in certain spaces.”

Bing Maps: The Deprecation That Wasn’t (Quite)

The October Power BI feature summary kicks off with an important update about Bing Maps deprecation – or rather, a reversal of the deprecation plan. Microsoft initially planned to remove the Bing Maps visual button, but reality intervened.

As John explained: “Too many people have to still use it. The data centers like Korea, government clouds, et cetera.”

The updated approach? They’re not ripping out the ability to have Bing Maps after all. They were only going to remove the button, and as Jason demonstrated, you could still work around it by opening an old report, copying the visual, and pasting it into a new one.

The takeaway: Bing Maps is still slated for deprecation eventually, but Microsoft recognizes they need to get Azure Maps supported in those unsupported regions before they can complete the transition. Continue updating to Azure Maps to stay future-ready, but you’ve got more time than previously thought.

Copilot DAX Queries: Finally GA

One of the bigger GA announcements for October: Copilot with the ability to write DAX queries in DAX Query view is now generally available. This feature has been in preview for a while, so the GA status means everyone can use it confidently.

John’s take on the value proposition: “People might quibble about its value as a production tool. I think it’s a great educational tool.”

The catch? You still need a Copilot-enabled capacity – meaning you need to be on an F-series SKU (or P-SKU if you’re still on Premium). For Power BI Pro users, this remains out of reach, which limits its accessibility for a significant portion of the user base.

Jason’s observation captures the frustration: “It’s just another teaser to get onto at least an F two.”

John floated an interesting idea: “Wouldn’t it be nice if they had some kind of a backend capacity for desktop to be able to do that sort of light stuff?” The hosts discussed how Pro users are on a shared P2 (equivalent to F128), so theoretically Microsoft could enable Copilot features for Pro users, even if they raised the price by a dollar per seat. But enterprise agreements and COGS considerations probably make this complicated.

Button Slicer Goes GA with Cross-Highlighting

The button slicer feature has graduated to general availability, bringing cross-highlighting capabilities that make for much more intuitive reports. Patrick (from the Cube series) apparently did a nice video demonstration of what you can do with images and creative implementations.

John’s assessment: “It just makes for some pretty intuitive reports.”

Also GA: the auto grid feature, which Jason described as “snazzy” and makes report layout much easier. His verdict on turning auto grid off? “Just as ugly. I wouldn’t even want to do it. It’s the old way. I’m glad that it’s there now.”

Visual Calcs Come to Embedded Scenarios

For ISV scenarios where customers embed Power BI reports, visual calcs are now supported in preview. This addresses a gap where embedded customers were being left behind – whether they’re using A SKUs or F SKUs doesn’t matter, they now have access to this capability.

As John noted: “It’s good to not leave those embedded folks behind.”

Table Columns: Auto-Expand Gets Mixed Reviews

A new feature that’s generating controversy in the community: automatically expanding table columns to fill available space. This one has gotten “some negativity” according to Jason, who’s been reading Reddit comments and the blog post feedback.

The intent is good – addressing a long-standing frustration with table visuals. As John explained: “You had to really anticipate exactly how it would be consumed and get your column widths right the first time. You start adding columns and it didn’t do it, and once you explicitly change the width of a column would always stay there and it was hard to get them all lined up.”

The important caveat buried at the bottom of the documentation: “This will not interfere with any adjustments you make to column widths. Once a column width is manually adjusted, it won’t be automatically expanded.”

John suspects this is a first attempt that will get tweaked based on feedback: “That team does listen to feedback.” But the vitriol in some community reactions surprised even Jason: “I get complainy sometimes and poke fun at things, but there’s legit vitriol around some of this.”

The New Power BI Controller for PowerPoint

When you embed Power BI reports using the native functionality in PowerPoint (it’s native now, not an add-in), there’s a new add-in available to help manage all your embedded reports in a deck.

As John explained the use case: “If you’ve got a bunch of different reports embedded in a single deck and you change something about them, whatever, a given property, and you need to do that across the board, that’s tedious. So this would be a way to basically apply a single change across all of your embedded reports in a PowerPoint document.”

Jason’s main gripe? “I hate that it’s an add-in. I wish this were just a part of another tab in the Power BI side of things.” Separate add-ins mean updates come in differently and create additional management overhead.

The real value proposition: If you’re creating decks for multiple customers and need to quickly swap which workspace and which report you’re pointing to, this controller makes that process much simpler.

Performance Analyzer Finally Hits the Web

The Performance Analyzer tool that’s been available in Power BI Desktop for quite a while is now available in the web service. This is another step in the march toward full feature parity between Desktop and the service.

Jason’s note of caution: “This is not something for the default user necessarily from a performance analyzer perspective, but for your more pro user, this is a really great thing.”

Export Power Query to Data Flow Gen2

An interesting data connectivity feature that caught John’s attention: the ability to take what you’ve built in Power BI Desktop and export/publish it directly to a data flow Gen2.

The workflow: Hit “Export query results,” pick a data destination through the Fabric UI, connect it, and the system will create the data flow Gen2 and run it to populate your destination – all from Desktop.

The old way required creating a new data flow Gen2, going into the advanced editor in Desktop, copying the code, pasting it into the advanced editor of the new data flow, and then wiring up the output. This streamlines that entire process.

Arm Support for Power BI Desktop (Sort Of)

Power BI Desktop is now supposed to be supported on Arm, which has John genuinely excited. But here’s the problem: the hosts spent 30 minutes trying to figure out how to actually get it working and couldn’t.

Both run on Mac using Parallels, and they have the September cumulative update installed, but it’s still showing up as the 64-bit version. There’s nothing in the blog post with detailed instructions on how to actually obtain the Arm version.

John’s commitment: “I’m going to dig in a little bit myself. Maybe by the next time we record, we’ll have an answer on that.”

The potential payoff is significant. When running Windows on Arm in Parallels on a Mac, you’re currently running an X64 version of Power BI Desktop that’s emulated. Windows on Arm is emulating X64, and Parallels is emulating Windows – that’s double emulation. Getting native Arm support would eliminate one performance hit.

As John noted: “It really works well, I got to say. But yeah, I want to get this up and running.”

Microsoft Fabric October Updates

Shifting to the Fabric platform updates, October brought several foundational improvements ahead of the big Ignite announcements.

UI Enhancements: Keyboard Shortcuts and Focus Mode

The new horizontal tabs UI now supports keyboard shortcuts – use ALT + number on Windows or Control + number on Mac to navigate between tabs. Helpfully, the documentation provides both Windows and Mac controls.

There’s also a new focus mode that removes all the Fabric chrome to let you concentrate on what you’re editing. Jason hasn’t found the need yet (“I’m still jumping between tabs so much that the busyness hasn’t bothered me”), but acknowledged it’ll be valuable when really diving deep into writing code.

Security Features: Outbound Protection and Private Links

Two security features moved to GA in October:

• Outbound access protection for Spark – Apply rules to prevent data from flowing to unauthorized destinations
• Workspace-level private links – Prevent external access without multi-layer firewalls

Data Engineering: Major Performance Wins

The data engineering updates are where October really shines. Two features in particular caught the hosts’ attention for their potential performance impact.

Adaptive Target File Size

When working with Delta Lake (Parquet data files), file size matters for query performance. You don’t want to open more files than necessary to perform a query. The optimal setting for a small dataset is very different than for a large dataset.

Previously, you’d have to manually maintain file size settings and perform reorganizations when they needed to change. Now the system can detect the best setting and apply it automatically.

The impressive part? Microsoft claims this delivers 1.9x faster performance.

John’s reaction: “I would want this to be default, frankly. I’m not sure why it isn’t. Probably because it’s a new feature and they want to make sure it works really well first, but I’m going to go turn this on.”

Jason agreed: “Why would you not just make this default?”

The answer probably lies in supporting edge cases. As Jason speculated: “For someone who is a Spark admin, they probably have reasons why they might want to consider keeping things a different way.” But for most users, this should be enabled.

Auto Compaction

Compaction removes old expired data and handles deleted records. Delta Lake includes historical data for time travel, which is nice from a recovery standpoint but not great for efficiency. When adaptive file size changes files, existing files won’t automatically change either.

Previously, you’d need scheduled processes to run compaction. Now the system can detect when compaction would be beneficial and run it automatically.

John’s assessment mirrors his take on adaptive file size: “Same exact thing – I think I want this on by default.”

New Spark Connector for SQL Databases

A preview of a new Spark connector enables reading and writing to SQL databases – including Azure SQL, managed instances, SQL on Azure VMs, and Fabric SQL databases. Notably absent: on-premises SQL Server, which suggests this won’t go through the gateway.

Jason revealed they’re already using this at work, and it’s made a significant difference: “This came out, I sent it over to the developer who was working on it, and Dallas was like, ‘oh wow, that worked really, really well.’ This just solved so many problems for us so quickly.”

The key improvement: seamless read and write capabilities. Previously they tried nine different approaches with various workarounds. The new connector made authentication much simpler and dramatically improved read/write times.

As Jason put it: “The longest thing we have is spinning up Spark pools at this point. The actual accessing the data…this made it rather seamless. You grant the right access and it flies.”

The connector supports PySpark, so you can weave it into notebooks naturally. The logging and authentication support also got noticeably better.

Additional Fabric Updates Worth Noting

Lakehouse Explorer Deep Links – Send links to specific tables in a Lakehouse instead of just the Lakehouse itself

Data Agent Integration – Add data agents directly from the Lakehouse with a simple button, rather than creating an agent and reaching back

Data Agent Authoring Experience – New UI with dedicated tabs for Data and Setup, plus markdown editor support for agent instructions

JSON-L File Support – Read and ingest JSON-L files using open row set in data warehouse (neither host was familiar with this format)

Graph Explorer Preview – Native graph querying capability, though the hosts elected not to dive deep since November brought more significant graph-related updates

Event Stream Updates:

• MongoDB CDC support
• Schema support from Event Hub
• Pause and resume for derived streams

Data Factory Enhancements:

• Copy job support for Excel, Avro, and XML formats
• CSV format improvements (quote characters, escape characters, encoding options)
• Variable library support in data flows Gen2

The Real Story: Setting Up for Ignite

As both hosts acknowledged, October was intentionally lighter because Microsoft was gearing up for Ignite. Jason counted “no less than about eight bazillion blog posts” in the week leading up to their recording.

The excitement – and controversy – around Ignite announcements is clearly top of mind. Jason’s description of Ignite’s purpose: “Ignite really is Microsoft’s vision moment where they get to talk about what they’re looking at for the next 18 to 24 months and to tell you what things they just released as general availability.”

The upcoming Ignite episode with Steph Bruno promises to deliver the hosts’ “brutally honest” takes on the announcements, including discussion of community reactions that ranged from excitement to visceral negativity.

Bottom Line

October 2025 delivered solid foundational improvements across both Power BI and Fabric, even if the release was lighter than typical months. The performance enhancements in Fabric’s data engineering stack are particularly noteworthy, with adaptive file sizing and auto compaction potentially delivering significant query improvements.

Power BI continued its march toward feature parity between Desktop and service, brought several preview features to GA, and made some UI improvements that sparked spirited community debate.

But the real action clearly happened in November with Ignite. The hosts’ promise of honest, unfiltered takes on those announcements – including addressing community controversies – makes their upcoming episodes must-listen content for anyone following the Microsoft BI and analytics space.

As Jason wrapped up: “Come back, listen to the November episodes once we release them. There’s good stuff happening there. You’ll also be able to get our honest take on some of the things that we saw.”

Links & Resources

• Episode 309 – Power BI September 2025 Feature Summary: GA Everything and Time Gets Intelligent
• Power BI October 2025 Feature Summary
• Microsoft Fabric October 2025 Feature Summary
• BIFocal Podcast Website

Subscribe: SoundCloud | iTunes | Spotify | TuneIn | Amazon Music