The Data Stack Show - The PRQL: Can One Table Rule Them All?

Episode Date: June 3, 2022

Eric and Kostas preview their upcoming episode with Ahmed Elsamadisi of Narrator AI.

Welcome to the Data Stack Show prequel, where we talk about the show that we just recorded to give you a teaser. Costos. If I was going to tell you, you have no budget concerns, you can get any sort of data warehouse storage layer that you want,

and you can do whatever you want with it, but you only get 11 columns in one single table. What is your gut reaction to that? Trying to remove yourself from the conversation we just had. I would be

very, very curious to see what this table looks like. And I would be very, very curious to see what this table looks like. And I would love to see it instantiated with data from a real use case to see exactly how it behaves and how it performs on a database or a data warehouse. So two things there I would be really interested to. One is how these things scale, scales as more and more data is added there. And how expressive it is, like what are the things that we can represent there with these 11 columns? And yeah, I would, these are like the two things

that I would be like very, very interested to see and learn more about. Well, I would, these are like the two things that I would be like very, very interested to see and learn more about. Well, I mean, the good thing is we got to learn about this actually and more. So Narrator, we just recorded a show with Ahmed from Narrator. And it's a fascinating tool that takes all of the data in your warehouse, purportedly eliminates the need for sort of joins on foreign keys and allows you to map activities to 11 columns, only 11 columns, which is fascinating, and answer questions really, really quickly. And it was an amazing conversation. We talked about, you know, how do you deal with an 11 column table? What are the performance implications? We also talked about turning everything into an activity and sort of living in a world where a time series table is the only way that you view data.

But Kostas, what were some of the things that jumped out to you in the conversation from a technical standpoint in terms of the way that they're managing some of the trade-offs? Yeah, I think it's... I mean, not only what Ahmed is doing is one of these very opinionated solutions where you actually abstract, let's say, the complexity of tasks so much that at the end, the biggest, let's say, obstacle or the trade-off that you have is how people will change the way that they think and they work with their data. And that's something that I think people that will listen to the episode, they will see how big of a challenge it is for a narrator to realize their vision. It's not just like, okay, we have the solution here, but like, you also have to sit down and like start thinking like in a different way to do that.

And that's like the main, the main trade-off there, right? Doesn't have necessarily like a technical solution to it. There are like things that you have to do where they are related, like to how to educate the users or like how to convince the users and how to create, let's say, processes to onboard them faster and like reduce the time that it takes for them like to stop thinking in this way. So I think that's like the trade though.

Like whenever we add like so much abstraction with technology, like we reach a point where like people will also start like to start thinking like in a, in a different way. It's not always... If you think about the iPhone, the first iPhone that was... 2007. Yeah.

You have a completely different way of interacting with this device compared to the past, right? Like with phones. But what was so magical about Apple is how intuitive it was to start doing that, right? Now, replicating this, it's a very hard task. That's why you have a company out there that's called Apple. I don't know.

It's a valuable company in the world. Yeah, exactly. Now, if our friend, I met like managed to do that, probably he's going to be the next Apple. I don't know. We'll see. But not like the main, let's say like trade over by Cedar and like the biggest like challenge that I also recognize there and both from a technical and a business perspective.

I agree. Well, I love the big thinking and you're going to love this episode. If you want to hear a very opinionated take on the way that we interact with data in the future, subscribe if you haven't and you will get notified of the next episode with Ahmed of Narrator

and we'll catch you on the next one.