C# Scripting in TE vs TMDL View – Ep. 396

Mike and Tommy compare the classic Tabular Editor workflow—C# scripting, macros, and model metadata automation—with the newer TMDL-based experience showing up in PBIP and Power BI’s TMDL view. They break down where TMDL makes collaboration and source control dramatically better, and where Tabular Editor still earns its place in a serious semantic model toolbelt.

News & Announcements

Introducing enhanced conversation with Microsoft Fabric Copilot and AI Skill (Preview) — Microsoft shared updates to how Fabric Copilot (and AI Skill chat) can store prompts and chat history to improve response quality and keep better context across conversations. The post also highlights a tenant setting for cross-geo data storage that can be required outside the US/EU, and clarifies that the stored conversation history isn’t used to train Microsoft models (and can be deleted by users).
Microsoft Fabric Copilot to write DAX queries in Power BI update — Copilot for DAX Query View can now use more semantic-model context like descriptions, synonyms, and column sample values to write better DAX queries. The big takeaway: investing time in model documentation (descriptions/synonyms) now directly improves both human usability and Copilot accuracy.
Submit a topic idea / mailbag — Have a question you want covered on the show? Drop it in the mailbag form—episodes are best when they start with real-world scenarios.
Subscribe to the podcast — One hub page to catch the live stream and find Spotify/Apple links to listen later.
Tips+ Theme Generator — Generate consistent Power BI themes quickly so your team can stop hand-tweaking colors and fonts across reports.

Main Discussion: C# Scripting in Tabular Editor vs TMDL View

With TMDL showing up more and more in the Power BI developer workflow (PBIP projects, Git integration, and the new TMDL view in Desktop), Mike and Tommy step back and ask the practical question: when should you still reach for Tabular Editor + C# scripts, and when does a TMDL-first workflow win?

Why TMDL changes the “source control” conversation

TMDL (Tabular Model Definition Language) pushes semantic model definitions into a clean, text-based, folder structure. That seemingly simple change unlocks a bunch of developer-grade benefits:

Readable diffs and reviewable changes (instead of opaque model updates)
Better merge conflict behavior when multiple people touch a model
A path toward “treat the model like code” workflows in VS Code and pull requests

Microsoft also calls out that Fabric Git integration is expected to shift from exporting TMSL to exporting TMDL over time—so learning TMDL now is an investment that should pay off as the ecosystem stabilizes.

Announcing general availability of Tabular Model Definition Language (TMDL) — TMDL is now GA, with continued investments planned for VS Code tooling, serialization control, and eventually TMDL commands (similar to TMSL commands) that can be executed against the XMLA endpoint.

Where Tabular Editor still shines

Even with TMDL improving the “model as files” experience, Tabular Editor remains a powerhouse for semantic model development—especially when you want to automate repetitive metadata tasks.

In the episode, Mike and Tommy emphasize that Tabular Editor’s strengths show up when you need to:

Generate or refactor at scale (measures, calculation groups, display folders, format strings)
Apply consistent patterns across many tables/columns/measures
Run scripting-based automation with guardrails and repeatability

Practically speaking, they frame it as: TMDL improves collaboration and transparency, while Tabular Editor improves speed and bulk-edit capability.

Workflow framing: “tools, files, and the skill split”

One of the more useful parts of the conversation is how they frame the human side of these tools:

TMDL makes it more approachable for teams that already live in git, PRs, and code reviews.
Tabular Editor scripting rewards people comfortable with object models and programmatic refactoring.

The likely future is not “either/or”—it’s a workflow where teams can use both, depending on the task and the skill set of the person making the change.

Bonus: scripting beyond the model (report layer)

The broader theme of the episode is automation—and that includes what’s happening on the report side too. Tommy highlights that the direction of Power BI developer mode and file-based projects creates opportunities to script “things that used to be click-only.”

C# Scripting the report layer with Tabular Editor — A deep dive into leveraging the newer PBIR report format (developer mode) plus C# to programmatically create/modify report elements. It’s an early signal of where Power BI projects are headed: more granular files, more schema stability, and more automation potential (with the usual preview caveats).

Looking Forward

TMDL going GA is a big signal: Power BI semantic models are moving toward a more developer-native workflow—where changes are reviewable, testable, and versionable. But as Mike and Tommy point out, there’s still plenty of room for Tabular Editor in the toolchain—especially for the high-leverage, automation-heavy work that teams do when they’re standardizing modeling patterns.

If you’re building a team workflow today, the practical play is to start getting comfortable with both: use TMDL/PBIP to level up collaboration in git, and use Tabular Editor scripting to move fast (and safely) when you need bulk changes.

Episode Transcript

0:32 good morning everyone and welcome back to the explicit measur podcast with Mike and Tommy hello good morning oh good morning Mike how you doing well have you ever had those weeks where you underestimated how much work it was for things I feel like I do that all the time but is it like every day for you it feels like that sometimes well this week I was trying to rebuild some of my existing content for training pieces which it all went well training’s gone well I have some the last is today so I’m got the content built it’s done

1:03 so I’m got the content built it’s done it’s ready to go I think it went well but man I thought I had it more together than I did and I was like yeah I’m good I got a lot of things prepared and I’ve been up like late every night just preparing and cleaning and making new slides and like H it’s been very good very rewarding but I just I’m exhausted dude I’m sorry to hear that I I know what you mean there’s a lot of coffee in my house right now in my office that I have myess machine and I was saying it’s a good and bad thing that it’s

1:33 a good and bad thing that it’s here so yes I agree with you so let me let me let me I have my coffee machine here as well at my house so mine’s sadly upstairs I have to go upstairs to go get it but before we do that so we’re talk about news and things let’s just give you the main topic for today our main topic today is going to talk about C scripting in tabular editor te as we like to abbreviate the names versus using tmdl or timel View and I’m actually going to throw in a a little bit of Dax quer review as well because I’m loving these

2:03 review as well because I’m loving these two tools that are now solely inside desktop so that’s our main topic for today let’s go talk about some news real quick here before we get going on the podcast all right Tomy what are you finding what’s what’s the news happening so we have some two great powerbi updates and curious with the blog do you think they’re going to combine the two and the fabric blog and powerbi blog into one because I’m going to a lot of different places all right it’s confusing to have two of them honestly I get that they are separate tools and I

2:34 get that they are separate tools and I think even the community is pushing back a little bit I’ve just read some articles who just been like why are we trying to make fabrick everything that powerbi is but Microsoft just recently announced I thought I just saw a couple tweets and or Microsoft earnings call that Microsoft powerbi has 30 million monthly viewer users every month I was like Wow Tommy we chose the right career to be in this in a growing space where they have that many users who need help with things that’s like that’s a pretty substantial number dude

3:06 that’s a pretty substantial number dude that’s we have a lot of more people to reach that’s what that’s what that means well this is me even years ago when we when I was just starting to do teaching about stuff I always I kept forget I kept thinking like everyone should know powerbi I’ve been doing it everyone else should know it but I keep remembering like powerbi has been out for like what 10 years now and now people are still just getting started with powerbi right now I can’t I couldn’t imagine just starting with powerb right now I’m looking at it I’m doing this class I’m teaching right now there’s a lot of things to learn like Dax where to write

3:36 things to learn like Dax where to write things I have Dax quer viiew I’ve updated my this is one of the reasons why I updated my training material I wanted to include timle editor and Dax query view because these are really good experiences for like automating and making multiple changes all at once way different than it used to be I think we we’re the tool is getting very very very powerful nowadays no I 100% and so I think with that number just means that we need to expand our audience so if if everyone knows five people that’s the Kevin Bacon

4:07 knows five people that’s the Kevin Bacon effect yeah and then we’ll eventually get to everyone knowing the powerbi our wonderful podcast that nobody listens to we probably have like less than one than 0. 1% of the population listening to the podcast at this point so tell your friends about the explicit M podcast this is talking about one of the most prolific and popular bi tools in the world so let your friends know that this is a good waste of your time for an hour exactly exactly exactly so my friends the first article

4:38 exactly so my friends the first article we had speaking of which is has to do with co-pilot and all the co-pilot Updates this is introducing enhanced conversations Microsoft co-pilot and of course slash preview any Improvement is what a better improvement over what you had I even have my shirt on today does co-pilot told me to say it say it yeah great choice how fitting how fitting so this actually has to do with

5:09 fitting so this actually has to do with just a better functionality for accuracy and the chat and context and I think you’re seeing all these other tools with context becoming more and more of a bigger thing sure that simply with co-pilot you’re going to be able to enable this additional context so not a whole ton on features and seeing it yet they’re just that it’s on the way okay but I think it’s the like Microsoft now has on their AI Foundry like deep seek and all everything else so they’re like oh I

5:40 all everything else so they’re like oh I wonder if we can integrate that so more context for my data please I will I will say this you please I will I will say this copilot is interesting of the know copilot is interesting of the places that I like to use copilot I’m going to talk about copilot as a general item I think co-pilot is deceiving a little bit because it’s like there’s a powerbi co-pilot there’s like the fabric co-pilot there is a co-pilot for office like they’re all they’re all co-pilot but obviously the code is not the same for all of them right it has to be more contextually aware of like what’s going on in the places that I find co-pilot is

6:12 on in the places that I find co-pilot is most effective for me it’s has been always inside the notebooks writing code co-pilots are really good and I really enjoy that part of the experience so anyways I’ll just point out like they’re talking about co-pilot for data engineering and data science yeah 100% I when I’m writing code or asking it how do I write a python function can create a a class how do I create a class it’s spoton it’s really good so that in that regard I I really like copil still jury’s out about everything else with co-pilot and powerbi I haven’t I’ve

6:43 co-pilot and powerbi I haven’t I’ve tried a couple times it wasn’t too impressive it did some things that were helpful but and where I think what open AI just released what is it their their actions so they have they have two things they have their deep research or their which is basically or the O3 model which is just basically that like the reasoning and thinking right but then the other thing is the operator you might be talking about that’s what I’m talking about okay

7:13 about that’s what I’m talking about okay so I think for me my my personal preference here is I I was calling this about a year ago saying like AI is good but just giving me like reasoning and thinking like that’s somewhat helpful I can use it to start my ideas about a topic or something I want to write about where I find AI has been really useful for me is in the generative space right generating images backgrounds things are not copyright protected like that’s huge for me so having stuff like that is massively improving what I work with but when I start having the agents or

7:46 but when I start having the agents or the the co-pilot when I can tell co-pilot to align all my visuals when I can tell co-pilot to switch out all the pie charts for bar charts and remove the title axes like right that’s to me that when the action part of it when the operator version of co-pilot comes out that we can start operating on reports and having it do multiple actions on things I think that’s where I’m going to really start finding value from co-pilot so I don’t I don’t think we’re there yet I I’m really hoping that that’s where we eventually land but just as a an outsider looking at that’s where I’d

8:16 outsider looking at that’s where I’d like to see it go no I I’m a thousand per with you just make that integration if it was as seamless as the GitHub co-pilot was oh my goodness would that be incredible for us right because you use the copil fabric and it’s fine it’s fine but compared to actually using the like the GitHub integration we like hey I can take any repositor I can take any package it doesn’t matter what it is

8:46 any package it doesn’t matter what it is give it to the co-pilot in GitHub like a repo which repos and we’ve talked about this it’s like hey could you about this it’s like hey could basically take a look at this know basically take a look at this what’s broken what can I do how do I use it why can we do that if we have tinle context yeah yes and I think I I feel like I saw Alex power say something like co-pilot now has the ability to read descriptions of measures and names and names of measures and a little bit more context around what’s going on there I think I think that’s where we’re

9:16 there I think I think that’s where we’re going to go though like I think this is going to emphasize more of the point

9:20 going to emphasize more of the point that you need well documented models you need more context about what’s happening in the model and I think co-pilot particularly as a gets closer and closer to like now it can see maybe it can write see daxer viiew queries maybe it can see timle maybe it can see all the descriptions out of the measures and things that makes more sense it’s going to push us to make sure we when we certify a data set or a semantic model we’re going to be focusing more of our attention on getting the descriptions and everything populated into the model

9:51 and everything populated into the model the model will be the documentation a th% yeah so I can’t wait but and it’s funny because we always talk about co-pilot to from the developer point of view and I I guess that makes sense but they’re trying to build both sides of the coin like you have the AI skills I don’t know if you’ve touched that yet I have not I don’t I haven’t really spent a lot of time getting in there I again I busy on things I’ve been I’m up late training people getting used getting started with powerbi I’m not I’m not having a lot of extra time to go

10:21 not having a lot of extra time to go spend on agents I also don’t have a lot of customers in my world asking me to help them build agents right so is the demand there not that I see for my customers so if a customer wants me to help figure something out they think it’s valuable to them then maybe I’ll spend some more time on it but right now I don’t have time to I I’m still more just trying to get people started with powerbi getting them using the the regular powerbi interface and governing that part that’s where we’re struggling right now right now yeah we’ll get there well speaking of more co-pilot updates something that has

10:53 more co-pilot updates something that has to do with more code and this is the second update this is on the powerbi blog not the fabric Blog the other one that we’re were talking about even though it was powerbi was on the fabric blog this is Microsoft fabric co-pilot to write queries in powerbi update it almost sounds like a New York Times title too rather than AI generated this is really cool and this is from our good friend Zoe So unless she’s using a AI to generate

11:23 unless she’s using a AI to generate titles and said please make this read like a news headline or PR release I like this much better than then amazing change your world thing with our new settings update yeah exactly right so going back to what copilot does best is this is what I’m talking about yeah I just made the comment something about like this with this update right just scri skipping the object I knew I thought I heard it somewhere this is what it was descriptions are proper in the model synonyms are there and simple

11:53 the model synonyms are there and simple value sample values all those now can be fed into the co-pilot to help you get answers that makes sense like that’s the kind this this is what I was talking about yeah so this is this is pretty cool this is pretty awesome we and we just talked about this too in our announcements or we did our fabric draft about the descriptions and titles in one L catalog well holy crap you want to talk about generation you don’t need know stinking script and custom code just ask copilot to help you look at

12:25 just ask copilot to help you look at descriptions go through them also they also talked about synonyms too we’ve talked about a ton when it done and it’s like that’s a almost a daunting task for an individual to do to go through an entire model and it takes a lot of time for something you may not use co-pilot’s got it so you can actually tell co-pilot to make the synonyms exactly that’s that’s the kind synonyms exactly that’s that’s the stuff that we’re talking about this of stuff that we’re talking about this is we’re starting to get to the point where things are actionable

12:55 where things are actionable right to me it it also feels like a point where we’re imagine you wanted to do some documenting of your model you should be able to like pop open co-pilot and just say hey co-pilot docum my document my model by adding descriptions and synonyms for all my measures and columns and what it should come back with is like a menu a screen that says here’s all the things that I think you should should add I’ve already pre-written everything for you just read them over do they make sense is am I right in what I’ve assumed about this information that would be really

13:26 this information that would be really helpful that and that’s exactly what we have because Mike I’ve gone through descriptions and been that person where you they want to use the chat feature and or even like the natural language feature with having the proper information there and it it’s a lot of time to do if you want to get it right for something you may not be sure that it’s going to be widely used and oh so this is a great way that copile will come in again going back to her point it doesn’t necessarily have to

13:57 her point it doesn’t necessarily have to have the most perfect description off the bat for a lot of people as long something and then we can update we can prioritize it rather than looking at all blanks I I’ll be very clear here I was in I was showing people how to use some of the AI visuals the two AI visuals that I like to use that Microsoft has produced is the decomposition tree which I think is pretty fantastic it’s a it’s a goodlook visual and it makes with a little bit of clicking around it makes sense immediately what you’re doing with it like oh it’s decompressing or

14:27 like oh it’s decompressing or decomposing this single number into all the different parts of that number where do it break down great makes total sense the other one I was giving an example around was the Q& A the Q& A Visual and so I was showing the settings of the Q& A visual which is where you can like do all the extra things right add the synonyms add the the terms add the descriptions like all the other things you need to do to make the natural Q& A better for users and my

14:58 natural Q& A better for users and my guidance was like look you guys are just starting out in powerbi don’t worry about this right now right don’t use the Q& A visual because it just gives so many not different but it gives it gives such a a broad spectrum of results that I feel like when I’ve used Q& A in the past I’ve gotten frustrated because I in my mind had I know what I want and the AI wasn’t able to I wasn’t able to speak to the AI or type in words the right way to the AI to tell it how to get the visual back out right so we weren’t

15:28 visual back out right so we weren’t quite there yet so I said look hold off first get good at getting your data together your first stage of powerbi is just automate everything get some solid semantic models done get people using the tool teach them how to use the filter Paine right here’s the pain points we typically see people having start there once you have really good working models and they’re being used a lot now let’s introduce Q& A on top of them so to me it’s like not a phase one thing it’s like a phase two or three thing where I’m going to really dive in and use the Q& A

16:00 and use the Q& A visuals yeah I can’t wait to see where this goes and honestly this is the really cool stuff that I’m I’m happy with with co-pilot and this is all going to be in powerbi desktop too right if I’m not mistaken I believe the images they’re showing is powerbi desktop and all this okay so but the trick is like you have to have something so this is this is the episode we had a a couple weeks ago where Adam Saxon jumped in and got us

16:31 where Adam Saxon jumped in and got us got our hopes up with gy a cube and then immediately dashed them halfway through the episode where we were like oh co-pilot can be used on F2 SKS like on a smaller skew you can use it wherever you want because my my developers may want to use co-pilot like if we’re if we’re finding all this value in like adding descriptions and adding measures and synonyms like if it’s a timesaver now it’s worth my effort to use these tools use copal to help me save time like that’s that’s worth to me I guess the question would be is how much is how

17:02 question would be is how much is how much is your time Worth right at that point if you’re doing a lot of documentation on models is it worth 20 bucks a month are you saving a couple hours of your time using these things maybe we’re getting probably are you probably are but it would require like so it’s the idea of like if I’m making lots of semantic models then makes sense to use this but if you did one model and you’re done it doesn’t make sense to keep co- pilot on forever and ever because you don’t need it all the time oh my gosh and it to and also too if some of it’s just going through code

17:33 some of it’s just going through code right if if I’m just going through either the timle queries that’s really just text at that point no you wouldn’t have to connect to anything theoretically so theoretically if I’m looking you’re talking I don’t have to connect to anything what are you talking about what is I don’t have to connect to a fat like a fabric Model if I’m looking at timle okay really it’s all text right like I don’t need to have a connection to the fabric

18:03 need to have a connection to the fabric service and look at the semantic model that way all the information is available in the timle are you just talking about like what you’re trying to prompt the co-pilot with yeah so let’s say I wanted like give me a summary I’m looking at everything in Tindle and I have that available and I’m in power by desktop well I say hey add descript to everything well it doesn’t have it shouldn’t have to go to the fabric

18:35 shouldn’t have to go to the fabric service look at that model and grab the

18:37 service look at that model and grab the information if it’s already on desktop in timle and text in the text format right like there’s no what information would it need to get so that should what I’m trying to say is that request shouldn’t be that large on your For Your Capacity okay you’re saying something yeah yeah yeah yeah so because basically right you don’t have to go look at rows of data you’re only looking at the metadata yeah but you’re the copad doesn’t run on your machine there’s no the model regardless if it’s local the model doesn’t come down to desktop it doesn’t run there so you’re

19:08 desktop it doesn’t run there so you’re always going to use some version of compute on Microsoft where I think things maybe get a little more interesting here is what if someone lets you use timle like you have the desktop you can get timle down I don’t really understand your mean I don’t really understand your point here very much you’re going to have to connect to the service at some level you’re going to have to get the the model down anyways the co-pilot runs in the service regardless what I think we’re finding though is that Microsoft is getting more efficient on the models like deep seek came

19:39 the models like deep seek came out and now we have models that can run at a fraction of the cost right so yeah obviously Microsoft and open AI are going to pick up on that immediately and start ripping that off and start making models that are third of the price to run and train like this all this is going to do is just going to make the whole the whole economy more efficient and yes DC came out and like I thought it would take out like a a trillion dollars of capital that came out like it just you wiped out a ton of money from these big companies but I

20:10 money from these big companies but I think all this is going to do is going to add competition now and these big companies will now have to make their models faster more efficient and this is all this is doing is making co-pilots or AI agents or whatever you want to call them more like a commodity it’s going to be the same thing as everything else we use we take for granted right computers electricity water like all these things that you just consume so much of it because you it’s easy to get it that’s what AI is going to be doing AI is going to get so efficient and easy that everyone’s going to want to use it it’s going to be

20:40 to want to use it it’s going to be interesting this will be the next couple years I think are going to be really interesting to see what happens here any other topics we want to cover before we get into our main topic today I think we’re good I’m excited to talk about this Mike all right well I’m this this second topic here talking about like timle and how timle works with co-pilot is very interesting to me it also is our main part of the topic today so I’m going to focus a lot of our conversation on jumping into

21:10 lot of our conversation on jumping into tabular editor so tab editor is something that you use Tommy to it’s an advanced modeling tool for the pro developer I would argue and then we’re going to see where does timle fit and I’m also going to throw in here Dax cor riew how how do these different tools help us and did timle just OB solete tabular editor do we not need tabular editor anymore I old statement let’s unpack this Tommy give us some a overview of what we’re going to talk about for the main topic today yeah so really it’s just with all the

21:41 yeah so really it’s just with all the announcements and updates to timle that’s been going on I really I think it’s probably due time to actually compare the timle way of scripting and you because you’ve talked a ton about this in the automation compared to what what the only option was available which was tablet editor and macros and scripts C scripting and its finest now you can use them side by side but really they are different workflows

22:11 but really they are different workflows and you’re kind you’re dealing with data in a different way you’re dealing with different files so to speak one’s tablet editor 3 one’s still going through the metadata of a model of a Bim Tims all this new language and it’s much more structured against something you’d be using vs code you can open a timle project in power or taable editor totally fine but again when it comes to what I can actually do once I’m looking at code versus tabl

22:41 once I’m looking at code versus tabl editor really I think a few things we wanted to discuss today is this coming of timle you like where do you see it what do you see that its benefits are better than the previous Enterprise workflows that are out there where does tablet still fit in or where I guess timle would not necessarily compete with and then how do we see this coming into the future where do I need to know both and like how the skill

23:12 and like how the skill levels are do they split up the type of person so really how are they going to play together so I think it’s really due time to talk about automation with the two coding or tools that we now have available to us timle and preview and tablet editor three or or two as well but CP macros yeah I’m going to be honest with you Tommy I don’t like writing C just point blank like if you think about the accessibility of these two different tools like tmdl easy to read I

23:44 different tools like tmdl easy to read I can clearly understand what’s happening in the semantic model with timle so already to me I’m already putting in the camp of timle editor or timle view it’s already got a win in my book it’s timle editor is made for easy use with the semantic model and and where do I use a semantic model I’m using it in desktop I’m building things directly there I’m making changes to things I got to be honest I was I’ve been teaching this class this week and we’ve had dax’s quer riew for a while now but I am heavily

24:16 riew for a while now but I am heavily heavily jumping into the the the D query view I’m writing queries I’m actually finding I like the experience of teaching new users about dimensions and facts and making it easier for people to use that information to write out the measure in the evaluate statement now it’s again it’s a little bit different right but I can right click on an item in the in the the scen scen navigation and then automatically I get

24:46 navigation and then automatically I get the definition of the measure I can right click and have a dummy measure populated for me that I can go modify just change the name of things which is awesome but then I was moving measures between data models and if you if your movie measures between models that’s a pain it’s a big pain but now you can do it with a couple clicks you can script everything out from one from one table all the measures copy the code go to a different model paste the code and then with a one button press update it this is changing my workflow

25:17 it this is changing my workflow on how I do things with writing measures so that’s this interesting place that you start right off the bat right when it comes to like copying things over because that is available in the tablet editor 3 interface just it’s not we’ll say as straightforward you can you you can actually create a Dax script yes where that’s copy paste and then find room place if you have table name you can click on the measures and say script measures or script table right and and it does its version of like hey here’s how to define this but again my argument

25:48 how to define this but again my argument here is they’re not using yaml they’re not using like a an easy to read structure it’s it’s whatever Tabler editor’s language is but again I’ll argue if you if the let’s talk about this the 30 million users that are showing up to powerbi to use it every day from Microsoft’s earnings call right if we have 30 million users what do you think the portion of those users who know C are I would argue most of this is business users so that’s a very I don’t know 5% 1% I don’t know some small

26:19 know 5% 1% I don’t know some small number right it’s not it’s not a line share of those 30 million users people are just showing up because they’re trying to build reports connect to us Excel files connect to things SharePoint get it into reports get it in power. com that’s the the majority of people I think are doing that work so if I look at that and say okay which tool is easier for that user Point Blank easier it’s going to be it’s going to be the timle view or the Dax query view it’s going to be easier I don’t want to write C so I think that to

26:50 don’t want to write C so I think that to me that’s just a big ding against tab Ed at this point not that it isn’t powerful listen yeah if you learn C and you can use it T editor yeah you’re going to save a lot of time 100% no no question there it’s just a matter of is the learning curve now too Steep and do I just want to use the actual desktop program and I think we’re getting to a place where desktop’s getting pretty good so let’s pause here because I I think the big focus is and I’m with right with you with the c scripts I

27:21 right with you with the c scripts I actually have always followed I want to give a shout out to Bernard ago who is maybe one of the coolest people in terms of writing CP scripts and I’ve tried diving into him and replicating what he’s done and like he he said he had to set up his own environment Visual Studio I’m like I I want to actually get really good at this because what that would be basically be able to do for me but it’s not easy especially if like again learning C is such a in a sense

27:54 learning C is such a in a sense Advanced language as it is it’s like

27:57 Advanced language as it is it’s like that a lot of the stuff that I’ve done is quick modifications there’s been a lot of times I’ve come on this podcast over the years where I’m was need deep into a c script and the podcast had to start and I’m like I just need to finish how to do this yeah and so yeah but a lot of people don’t have that time well just simple things like let’s let’s just say I’m gonna I want to script out all my measures from the table and just see what the formatting of every measure is just look at it do that in desktop right now you can’t do it that only appears years in timle

28:28 it that only appears years in timle editor and it’s literally again again it’s the point of like the which where’s the friction coming from right so on one hand if I am going to pay for tablet editor 3 I have to pay for it right on the other hand if I’m using tablet editor 2 it’s free right but now I have to go into desktop I have to click on external tool I have to get the the tool open right okay not super friction filled but those are like three or four extra clicks yeah now timle editor I just go to timle editor drag the table

28:59 just go to timle editor drag the table drop it in the window boom it just shows up ruie Ru is the PM for the Tim timle editing stuff and dude he’s killing it he’s he is a developer he knows what to build everything in that tool is just right it’s done correctly so good I can’t speak highly enough about it and it’s interesting you talk about two the like even some of the replication right and because I’m going to share something about my own tablet editor environment and again I love it I use magros and

29:30 again I love it I use magros and everything all the time but it’s important to note they’re not synced anywhere like I actually take my tablet or environment and basically have a g repo of it that I have a and again you want to talk about convoluted mic what I have to do because if I lost my macros I would I would lose myself I actually created a dang script and this was one of those like two in the morning things where it will actually look at the tablet edor macros copy them put them in a repo and then

30:01 copy them put them in a repo and then zip it up push it to a a private repo so just in case something were to happen I can always be in sync but that’s it’s very convoluted to do that because those macros I could not replicate because I don’t know C now that I have them they’re great but now to your point with timle you you got those scripts from like is that from Bernat that you got them from or you just found them from on the Internet or how’ you get them so one started as inspiration but this was my this was like a two-day three-day thing where I just didn’t check email

30:31 thing where I just didn’t check email and I’m like I just wanted to get this it’s actually really cool but yeah this was this was Tommy learning C which would meant Tommy was having a hard time learning C but it works but it’s still eluded because but I don’t understand you made these from you said you you made them some you you made them then I guess is what I’m asking yeah so someone had a script where they could just basically copy like macros from the like the local file I’m like I wonder if I could create something that would basically okay take all my settings yeah

31:04 basically okay take all my settings yeah configured in a zip file so I have the latest date backup thing and so then I can push that up and I can also use that same script if I go to another T like computer it will then push it to the like wherever TBL editor reads from so it’s cool seamless but again it’s the only thing that really table editor 3 is if you don’t have those macros right like having the macros it’s amazing the

31:34 like having the macros it’s amazing the workflow without them it makes it really hard now compare that let’s compare that to timle what you’re saying here with the the environment anyone can download vs code in that environment is pretty easy because that’s when I’m using timle Mike where are you going to recommend people to go powerbi desktop in that view or the T the okay okay let’s dive into that one like so here’s what here’s what’s happening desktop is eating into that Pro developer space a

32:04 eating into that Pro developer space a lot more than it used to be right these are things other I’ve heard the complaint from other companies right hey our company will allow us to use powerbi because Microsoft made it great we won’t let you use third party tools because it’s not sanctioned by Microsoft it’s not secure we don’t know who they are right there’s all kinds of reasons behind this right that’s not our policy whatever fine that meant all the people who were using powerbi had their hand strapped to agree it it was a little bit less efficient for them to automate or do automation or repetitive things right

32:34 do automation or repetitive things right I’m telling everyone right now you when I when I used to communicate about powerb desktop I used to think about two different windows I don’t really use the data view very much the table view table view is interesting but I don’t really use it very much it’s like looking at the information I couldn’t write dacks on it I couldn’t manipulate anything it didn’t really do much right so table view was in my opinion okay but I would talk about power VI desktop top as being here’s the there’s there’s two things the desktop builds it builds the report side and it

33:04 builds it builds the report side and it builds the semantic model that’s and we can see that when we publish things to the service so I’d always communicate to people look desktop is building a single PB file but it’s actually two components it’s the semantic model and the report before I would only communicate about the model view about this is how we look at and view the model that’s where we do most of our model work now well now we have Dax qu review and timle editor so now I have three Windows to help me design build shape look at the

33:35 design build shape look at the information that’s there it’s so much easier to read it’s just it’s much clearer everything is clearer and so and this also I think is an advantage for us as as powerbi developers any new feature Microsoft builds if they add something let’s let’s say they add this new cool feature inside the semantic model maybe they embed co-pilot as part of like I don’t know configuring co-pilot inside the semantic model as part of the semantic model I don’t know if they’re going to build right any new feature they come up with can automatically be accessed

34:06 with can automatically be accessed through the timle editor the feature is there I’m not waiting on I don’t have to wait on other tools like tab editor or yeah other like Bravo like they all these other tools you have to rely upon them to update themselves so they can use the latest features of the semantic model well because timle is built into the tool as new features are released Microsoft themselves will make sure that that part of the tool is always going to work with these latest semantic model features or if they’re building new

34:36 features or if they’re building new semantic model features they’re going to make sure that it works inside the timle editor because it’s part of their tool so for that reason I think like it’s it’s going to be I’m I’m I’m finding as me personally I think desktop is becoming more and more the pro developer tool and not a beginner tool I think you’re going to be able to start really pushing on people to build reports in the surface at a much higher rate I don’t think you’re going to need

35:07 rate I don’t think you’re going to need desktop as much anymore and again just all the things that I’ve seen from the product team almost all the things that are coming to desktop they’re almost all going to make it to the service as well which is interesting to me that they’re they’re building these features in desktop first like timle editor you can use that in desktop but you can’t use it the service there’s no timle editor equivalent in the service yet I have suspicions that that will probably be a thing at some point in the future so as it seems as we

35:37 future so as it seems as we transitioned into timle because again has the timle scripting it’s we know it’s going to be more collaborative I guess are you finding yourself really not opening Taver editor as much anymore I wasn’t opening a ton to begin with I used it a couple times for some certain things like when I like where was I using T editor right I’ve got a really old analysis Services model that I’ve got to move into powerbi right I’ve got to script it out and figure out how to make like that’s when I would use it I’d go after the Bim I could see the whole Bim everything was

36:07 could see the whole Bim everything was formatted very well great let’s imagine I wanted to to go to a semantic model and I wanted to get all the little tiny files out of it right I want to take the model and not have it just become a single Bim a Json file I want it broken out into small Json files that’s what I would use Tabler editor for but now a lot of those very basic features I don’t use it anymore unless you’re so in tab editor if you’re not not doing Advanced things like writing scripts and consuming scripts and heavy

36:39 scripts and consuming scripts and heavy automation I don’t what else are you using in it no what what are you doing in there that you can’t do in desktop so here’s here’s one of the arguments that I think I would make as much as the timle environment much more readable some people still like a user interface and I know you say powerbi desktop has one but let me let me make an argument here the ability to quickly edit items create items in a user interface is more efficient in table editor I would argue

37:10 efficient in table editor I would argue to quickly create a measure to quickly go through the properties I think

37:14 go through the properties I think that’s easier in TBL editor because it’s you’re you don’t have to load the whole model it will push it back I still have that nice user interface and I can edit multiple things with powerbi desktop like creating a measure is fine you have the formula bar and if I were to create a measure using Dax I’m gonna use the daxx quer box in table editor not powerbi desktop it’s yeah but that’s my point though my point is I used to do that like that

37:45 point is I used to do that like that used to be my My Method with Dax query view I don’t need that anymore like I can I can right click and Define a measure and it writes the whole measure for me all I have to do is update the me so take that best part that thing you were saying that was in tabular editor that now exists in desktop so like if I’m writing Dax now I don’t like to write Dax on the model view I don’t like to write Dax on the report view anymore because in my past building experiences I would use the report view I’d make the

38:15 I would use the report view I’d make the measure and then I would immediately go to the model and grab some details and throw on a t on the page the report page I would immediately throw down a dimension and throw the measure right in it great I don’t have to do that anymore I write the measure in dax’s query View and then right click on the measure and say Define and evaluate it will immediately make a summarized columns for that measure for me and then I get to start adding Dimensions however I want just start typing it in and that’s really efficient and when I’m coming back to that it doesn’t clutter up my

38:45 back to that it doesn’t clutter up my report space it doesn’t make extra pages that I don’t care about like my whole workflow is changing as we as we speak and I knew this was going to be a game changer because this is such a big change and it from a developer point of view I agree and I’m finding myself transitioned I will play the devil’s advocate here it’s so developer heavy it feels like we’re it’s basically looking at yaml right like more or less if you were to think of some equivalent Co like

39:15 were to think of some equivalent Co like a syntax I don’t understand your comment here because what you’re describing is the same thing that Tabo editor does it’s so developer heavy there’s not there’s not any more UI differences between the two like I don’t I don’t see your point there so I’m making the argument that tablet editor you don’t have to create the macros in order to in a sense start automating or be efficient I can download tablet editor desktop today I can use the debug view go through everything I can create multiple measures at a time if I do need

39:47 multiple measures at a time if I do need something where I want to create a bunch of measures on the Fly there I can just Google search find that repo and I’m pretty set thing to get the basic automation done I have not I’m really not looking at code all everything’s in a user interface yeah I’m going to agree with you but a couple of the friction points that I have with t editor is and I don’t again I’m not in it as much as I used to be right I’m doing more architectural and governance things as i’ I’ve gone through my career now but things that I find

40:17 now but things that I find that are more challenging in tap editor is I could never find which screen is the right screen to put up I’m always getting lost in where that view lives that old vs Studio yeah I don’t really love that it seems old it doesn’t seem fresh like other things now so one thing I will give a lot of credit to so I have been really poo pooing tab editor through this conversation so far but I will give it some good credit is Tabet editor does do a good job of exposing all the properties for every single thing you need right so the properties pain on the

40:47 need right so the properties pain on the right hand side when you click on a table partitions all the different things you can see everything that it’s making inside the the UI there so I think in that in that regard that is extremely helpful and seeing all the properties there that’s something that the timle editor or Dax ciew does not provide you there’s no like single window with all the properties which is fine but you may want to adjust a property that you don’t know the name of it and so for you to go in and dig and

41:17 it and so for you to go in and dig and find it it’s just a little bit more work so I would that that would be one area where I would say I’d give a a thumbs up or a nod to Tabo editor it does a good job of giving you all the things you can do with the model and gives you all the properties even if they’re not filled out yet timle editor doesn’t give you that it only gives you the properties that have been changed changed so I have a qu I don’t know if you had anything else on this because one of the things with knowing that we now have 50 million powerbi users 30 million sorry excuse me

41:49 users 30 million sorry excuse me don’t over inflate the number I’m not gonna over inflate we’ll we’ll get there we’ll get there we’ll get there do you see in terms of the really the primary audience for either I’m using tblo editor or I’m using timle is it the same person in terms of like that same Persona because I think you can make the argument that really with from The timle View that is really just developer heavy and I could just be a powerbi

42:21 heavy and I could just be a powerbi expert in used tablet editor they over yeah so where do you see it going in terms of both for these heavy we’ll call them I’m not going to say just developer tools but Enterprise tools or or Advanced tools I think I’m going to answer your question with just talking about like a spectrum of people who use things that are powerbi let’s talk let’s talk let’s start the ends right so when we talk about powerbi I think I’m looking at the Spectrum of someone who consumes a report all the way up to

42:52 who consumes a report all the way up to a data engineer barely using reports but just shaping and making data okay so I think if I’m going to make the goal posts those are the those are the different end zones or those those are the two different ends of the spectrum perfectly yeah perfectly Fair okay so if we if we walk from let’s let’s start with the conversation around a user who is only consuming reports right they’re only using apps they don’t need any of this stuff they don’t even need desktop for that matter they’re just going to go to the service they’re going to consume it all right so one layer deeper than

43:22 it all right so one layer deeper than that would be someone who’s building reports not necessarily making semantic models they’re given a semantic model and they’re just making their own report pages okay they’re going to probably use the service maybe they’re going to use some powerbi desktop they’ll probably make some Thin measures very easy ones albeit inside the the thin report but nothing too heavy maybe they’re going to make some Dax inside a visual calculation although I don’t actually ever use that feature I don’t really see it valuable but okay some people will do

43:53 it valuable but okay some people will do it let’s move further to the next Persona which is the data modeler the data modeler is like a really good blend between someone who knows how to build reports and also knows how to build data engineering things they need to be technical enough to write and read code but they they don’t need to be they need to also have an an understanding of the business logic how do I shape the data tables right so it’s it’s I think the data modeler is a very senior business analyst right it’s it’s the one who’s been doing like not a brand new one not fresh out of college

44:24 brand new one not fresh out of college right they they know what they’re doing they’re more seasoned and they’re able to show up and really build that semantic model that makes sense right and then we start moving towards data engineer and I would probably say from the data modeler we talk data engineering right okay so you asked me where do these tools fit where do these tools align I think both timle View and Tabler editor 3 are sitting on that very technical end it’s the data engineer level person or the very Advanced semantic model people those are the

44:55 semantic model people those are the individuals that I think will be using these these tools so is one tool better than the other ones yeah tablet editor I think has been around longer it’s got more development time on it but also is more expensive right again why is powerbi so well adopted well it was easy to spr land and spread in organizations because a lot of it was initially free and then people started finding value from it and then people started paying for it so I think if you if you compare the price point of like tabular editor and timle

45:27 like tabular editor and timle Tim’s going to win every day and so I’m okay not spending some extra time learning C scripts if I can just take a little bit longer in a free tool that I can use and get around in so that’s where I’m at right now I think I don’t know if that really answered your question but I I I think these two tools address that really high-end data modeler and data engineer those are the individuals that are going to be using these tools that are going to be building large big

45:57 are going to be building large big semantic models and I would also argue too the larger your data model gets in size the bigger amount of gigabytes it takes the more you’re you’re going to want to use tabular editor because it’s going to make some of those larger model experiences easier to work with and I’m thinking specifically like refreshing your model building partitions creating incremental refresh some of those things are hard to diagnose and see what’s going on and that’s where the tool really shines I think yeah and honestly I I think think

46:27 think yeah and honestly I I think think from the heavy developer view like timle is going to be the way to go it’s it’s

46:33 is going to be the way to go it’s it’s going to be more familiar probably with any other projects you’ve done but I still want to I think still make the case that there’s still a road where I’m not a developer I’m just a heavy powerbi user and there’s still the user interface part that’s nice that I may not ever have to put my feet into that side of the water looking at the code I probably eventually will like I and I see that happening where I’m with you I’m going to disagree with you if you’re if you’re a heavy power B user you better be like I’m not hiring you if you

47:04 better be like I’m not hiring you if you don’t know how to use the code honestly seriously if you’re if you’re a heavy if you tell me I’m a heavy powerbi user and you haven’t spent some time learning like what the Bim is getting into timle doing those things I’m sorry I’m not interested in hiring you so like heavy powerbi user I I solely disagree with you I don’t think you I don’t even I don’t think I would require you to go learn tabal editor if you have those skills great I’ll take them great you skills great I’ll take them great how you you understand what’s going

47:34 know how you you understand what’s going on but like I’m not going to require an advanced power user to to require them to know how to use tablet editor I think honestly I would be more advantaged to hiring you if you already knew how to use the Bim the timle and the Dax query viiew to your advantage interesting so I is I if you’re Advanced you better learn how to know you better know how to write code code should be your thing if you’re

48:05 code code should be your thing if you’re Advanced any and any like doesn’t matter oh that’s that’s such a hot take I feel like in our a beginner user is the one who needs user interfaces a beginner user doesn’t know what properties exist right that’s that’s for beginners oh I this is a hot take Mike because how many powerbi Pros do who are are Advanced Stacks but they don’t like they have never worked on a python project whatever the case may be that python you’re whatever the modeler like

48:37 python you’re whatever the modeler like you better know like you better understand like how the tables relate like what properties are being set on the column again it’s a semantic model it’s all metadata it describes what you’re trying to do with the model right if I have a better user interface with timle which again timle is hugely beneficial there’s tons of value from this like that’s that’s what I want I’d rather so let me let me say this way let me just give you a comparison if I had a 100 hours of time to train people on

49:07 100 hours of time to train people on right and I’m gonna say spend 100 hours working on tap or editor learning it or spend a 100 hours learning the Bim the model and the code that comes out of timle I would choose go learn that in my opinion because that is going to be long-term sustaining and we’re just in the beginning stages of timle editor we’re just starting who knows how good this is going to get in the future I this is going to get in the future ru’s on it ru’s got an amazing team mean ru’s on it ru’s got an amazing team he’s got great features we’re just starting the space of this we

49:39 just starting the space of this we did a live video we did a demo of it and Ru was talking about like we have create or replace right now as far as an action in the the timle editor Ru is also talking about other actions that are going to be coming out additional actions so that’s going to be huge when that gets to become more capable so I think to me like if I look at the two options here the runway for what timle can do I think is much more open right now and it’s coming directly from Microsoft let’s drive adoption I me I’m

50:11 Microsoft let’s drive adoption I me I’m this whole episode of me is like get in there learn the tool go click the buttons let’s get some analytics on this for ru’s team because if you’re not in there playing with the tool ruy can’t get funding to build more cool features for us like that’s I want that to happen this is what and ultimately though to the way timle is going I I do agree with you I’m obviously playing a bit of the devil’s advocate here but let’s get this GA man and I obviously there’s gonna be that I don’t know

50:42 there’s gonna be that I don’t know what’s gonna take but until it’s GA too that’s going to be another hindrance but it is what I’ve wanted all along to be able to look at my power by model and my report and and non it doesn’t have to be non-binary so to speak let let me give you one other lens here as well let me let me just blow your mind here for a second Tommy you’re a big proponent of co-pilots or or AI agents agents right if let me just let me just

51:13 right if let me just let me just hypothetically here right Microsoft if you’re listening please bring down my capacity to use co-pilot down to an F2 I need this this is a feature I need honestly cuz don’t I don’t have a ton of money at my company a lot of companies are just starting out with fabric we need co-pilot at a lower level so we can start seeing the value of it at a lower level now here’s my point Tommy if you could use co-pilot inside timle editor and giving it actions to build code for

51:44 and giving it actions to build code for you who cares about the UI honestly real right like if I can if I can chat with the co-pilot and highlight a bit of sex and say explain this what does this property mean format all my measures that are format all of my columns to have this to have this kind columns to have this to have this formatting I don’t need to know the of formatting I don’t need to know the syntax I don’t need to know it just knows what to do and says oh yeah I know what you’re talking about and it goes through because it’s going to be able to learn on top of hundreds of thousands of

52:14 learn on top of hundreds of thousands of other models it should be able to be programmed on all the syntax of every single formatting for every single kind single formatting for every single column or measure or visual like the of column or measure or visual like the the model pieces like at that point I’m like well screw it I don’t want to even have the UI I want to just see the code and I want to talk to a co-pilot and say do the thing and things that I don’t understand or no the co-pilot knows about them the co-pilot will suggest something for me and it will work so that’s let’s I’m going to say I want timle Editor to just totally

52:44 want timle Editor to just totally LeapFrog anything that has a UI with it just get rid of it just go right to just right to code that’s I feel it warms up on that take but everything up to that point I agreed with you because again and I we live in that world we love that like we are more comfortable being in a project like vs code or something like that than not but again you take the majority of people who are like I know Dax but now I have to all those other things too yeah

53:16 have to all those other things too yeah I I that’s where I will I divert from you everything else you said yeah no and like we’re completely going that way now taking away the UI that may be look you already want to take away desktop put it in the web now you don’t even want on the web I it’s just a theal just get rid of all of it just I don’t want to have anything that’s hilarious no it’s not it’s I we’re gonna have to have something to work on but like I I agree like I think I think I want more web versions of the stuff I want timle Ed to

53:47 versions of the stuff I want timle Ed to get into the web I want copad to be a bit more everywhere in here like that’s where I think this is going to start really speeding me up and and I I I think I think we really need to talk in terms of like what does that data modeler role look like I think Microsoft is basically Microsoft has two roles data analyst or data engineer and I think there’s a middle ground in between those where you have an advanced analyst but maybe not quite a data engineer and that data modeler person there’s a specific set of skills you need in data modeling about working with

54:19 need in data modeling about working with large models and things there just there’s just deeper skills that are required there that I don’t fit think fit in the analyst role and they don’t think they fully fit in the data engineering role at least traditional sense of data engineering all right with that I want to start wrapping here final thoughts if you’re not playing with timle editor definitely go do it you should go look at the timle view inside powerb desktop I believe it’s a preview feature you have to go to your desktop and turn it on to use it highly recommended I am finding a ton of value from it Tommy’s a little more

54:49 value from it Tommy’s a little more skeptical but we’ll get there with that I just want to say thank you all for listening to the podcast this was a good discussion a number of hours here just spent with us thank you for participating if you find this discussion interesting if you liked what you heard here please go share with somebody else and if you’re running or exercising which I’ve heard a lot of people say this this is like my walk of the day I listen to the podcast thank you for walking with us we appreciate it and good luck on your exercise Tommy where else can you find that podcast oh you can find the podcast on Apple

55:19 you can find the podcast on Apple Spotify wherever get your podcast make sure to subscribe and leave a rating it helps us out a ton and please share with friends since we do this for free do you have a question idea or a topic that you want us to talk about in a future episode head over to power tips podcast leave your name and a great question and finally join us live every Tuesday and Thursday a. m. Central and join the conversation and all of powerbi tips social media channels awesome thank you all so much and we’ll see you next time

55:50 see you next time [Music]

Thank You

Want to catch us live? Join every Tuesday and Thursday at 7:30 AM Central on YouTube and LinkedIn.

Got a question? Head to powerbi.tips/empodcast and submit your topic ideas.

Listen on Spotify, Apple Podcasts, or wherever you get your podcasts.