Jakarta EE JakartaOne Livestream 2022, 15 Dec 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Jakarta NOSQL: Powered by Cosmos DB On the Cloud - Jakarta One 2022

Description

Jakarta NoSQL is a new standard for accessing non-relational databases on the cloud. Cosmos DB is a best-of-breed planet scale NoSQL database on Azure that is compatible with MongoDB, Cassandra and Gremlin.

In this session we will see how to use these technologies together in cloud native Jakarta EE applications. Most of the session will be demos with a minimal number of slides.

-

Jakarta EE gives developers a comprehensive set of vendor-neutral, open specifications that are used for developing modern, cloud native Java applications from the ground up. With Jakarta EE, technology developers and consumers can be confident they have the best technologies for developing cloud native, mission-critical applications. And they can build on decades of Java developer expertise to move existing workloads to the cloud.

Learn more: https://jakarta.ee/

Follow us on Twitter: https://twitter.com/home
Follow us on LinkedIn: https://www.linkedin.com/showcase/jakartaee/
Like us on Facebook: https://www.facebook.com/JakartaEE/

A

A

All right, hello, everyone, oh I, hope you can see my screen all right and you can hear me all right. You can't please yell and uh you.

B

Know you cannot see your screen yeah.

A

um I haven't haven't shared my screen yet I'll do that in in just a moment, so um I believe you don't know me. My name is Reza Rahman I'm, a principal program manager for job on Azure I've, been in the ecosystem for a long time now, I think ever since Java E5 time frame, man I don't even want.

B

To begin to count how many years that's.

A

Been uh but yeah I'm, also a in my uh outside of my day. Job I also help lead the Jakarta ambassadors. It's a group I'm very proud to have founded, and the talk today actually really is more or less along the lines of my day. Job I don't know if you can even see the shirt, I'm purposely wearing the shirt, and it says you know Java Microsoft Azure. This is like my official uniform. If you will it's not good, it's okay, uh so a few quick things.

A

First, before we dive in uh first and foremost, uh you know, don't hold back questions until the end uh generally I, you know I, don't usually do that for any of my talks. You can just ask me questions throughout uh and you know I'll just answer them throughout there'll be natural pause points in the talk where I'll strap a little bit and you know see if you have any questions and I can answer them, then you know so don't bother waiting until the end.

A

Just just ask me questions throughout all right, so uh without further Ado, let me uh get things rolling, uh so I'm gonna just share my screen. I'm gonna put my camera to the side here. I may uh glance sideways a little bit. uh If I do that, you know what I'm doing I'm. Basically looking at my camera, making sure things don't look messed up. That's all so! Hopefully you can see my screen now uh and hopefully very soon.

A

You'll be also uh able to see my slide deck and this pesky thing doesn't want to go away. uh You probably can see it, but this is like hiding my my buttons uh but yeah all right. So hopefully you can see my screen by now uh so yeah. The the title of the talk, as you know from the agenda, is a nosql powered by Cosmos DB on the cloud. uh What is the talk actually about? uh Well, the talk is about actually uh sort of a mashup of two different things, as you can imagine.

A

uh So it's a mashup of Jakarta, no SQL I'll talk to you a little bit about what Jakarta SQL is in case. You don't know what that is, and basically how you use Jakarta know SQL using a technological, Cosmos, TV and again I won't blame you.

A

If you don't know what that is, uh you know I'll explain what Cosmos DB is and how to use these two things uh in a in a Jakarta application in hopefully in a cloud native chapter, 3 application where you, you would have a use case for something like no signal and actually I'll spend a little bit of time talking about the nosql landscape in general uh internally, when I deliver this stock, that's actually where most people are still have questions on, which is understandable. You know it's still uh unclear.

A

Let's say you know where, where precisely the use cases, it's it's a worthy discussion to be having all right. So at the end of the talk, I will uh share out where the slide deck is you don't have to wait for it? It's already on speaker deck and if you already know my speed speaker Tech account, you can go, find it out uh and in fact you know all the code. Examples and I do have a little bit of good examples.

A

For this talk, all those good examples are also on GitHub again, you know my GitHub account. It's really easy to figure out which, which repo I'm talking about all right. So without, let's move forward. So, first and foremost, we should talk a little bit about Jakarta SQL, so.

B

Hopefully, a lot.

A

Of you here already know what Jakarta SQL is, but in case you don't. Let me give you a very brief introduction. uh If you want. This is a bird's eye view. um You know my job is here not to explain jakartano SQL, the mechanics of that too too much uh mostly I, actually want to talk about the integration pieces between uh Cosmos division SQL, but I'll. Give you a little bit of an overview. So Jakarta Municipal is a specification.

A

Hasn't even had a final release, yet uh hopefully it will soon. uh I'll explain a little bit why it still doesn't have a final release yet, but hopefully it will. But honestly it's been around for a little while and I know at least a couple of folks that are using this in production. So it's definitely production ready uh and you'll definitely begin to begin to evaluate it, and you know, maybe even like those folks that I know actually begin to use it in in a production environment.

A

It's very handy if you have a use case for it, so its objective is to create a standard API for accessing all manners of almost SQL databases similar to what we today to take for granted with something like jpa or jdbc. uh Sql is trying to accomplish the same thing for the non-sql landscape and, honestly, that's a very, very big task. So it's not so simple, because, unlike this, the relational database landscape, you know SQL landscape. It's very, very diverse. There's never really been a standardization effort in the nosql landscape.

A

Every single database is a little bit different and it has different features and different apis uh and honestly, it's it's kind of a hotspot of many different things so trying to standardize and create one API, for that is a pretty daunting task, but I do think uh the Jakarta SQL specifications so far really has been, has done this in a fairly Smart Way.

A

So one of the key thing, the the one of the way there's a couple of different ways why this is possible, but the way this is possible is actually is not just one API, uh it's actually several apis and it has sort of flexibility built into it that allows for uh basically adapting to many different use cases. So this is why it's actually a useful facade for something like the ninja foreign landscape. So, first and foremost, there's several layers in this API, so uh the lowest level uh API is the communication API.

A

It was very similar to jdbc if you were to create sort of your own binding, projector SQL. This is where you would start. You would say: okay, you know this is the particular communication API that makes sense to me. I'm going to implement it, so this will talk my language, the language of my particular nosql database, uh again very similar to what jdbc does in the relational world. You know every single data designer has a jdbc provider.

A

Similarly, um you know each every single provider would have a communication API for Jakarta sql.cio, a layer above that is a mapping and template API, and this is very similar to basically jpa right. So this is a uh what to use what you use to say: okay, here's a job object. You know this is how it maps to my underlying database, whatever type of database.

A

It may be, and there's a template API to actually have operations of over those right, so crowd operations typically create, read, update and different variants of uh of uh queries and we'll talk in a moment. Not every single, no single database even supports queries, but some mechanism of querying. So that's what basically jpa does right.

A

The jpa Entity manager, the jpn notation, is basically that's what it does uh yet another uh abstraction layer that Municipal has is something called the repository API, and this is very similar to if you work with uh with spring data jpa or the Delta Spike data model. That's what it is. It's basically an interface interface that says hey. This is how I access my database.

A

This is my finders. This is my sort of my crowd methods and basically generate the rest of it. For me- and this was something that is built into the jakartano SQL uh API. It's a me personally, I think it's a bit excessive. It's too much abstraction for me. I tend to not use it.

A

So in my code example, I actually won't show you the repository API, but if you're a fan, definitely it's available and there for you, you know again, if you're familiar with something like spring data jpa, you know how that looks like uh I will actually just be showing you, the mapping and templating API, that's more. In some cases. For most single databases, the key can be quite elaborate and have its own metadata mechanism, and the value would be like whatever it is, that you're storing in the database.

A

Typically, this would be some kind of binary object or it could be, uh you know, simply a representation of data in some way uh you know. Typically, this would be stored in something called Json or bson or whatever of you it's a lot of these notable databases will have richer built-in type capabilities.

A

The other one is column family, and this was an interesting one. So.

B

A

Variant of key value pair, basically, so you still have keys, but the value is in the form of a set of columns themselves right. So, instead of just having a Blog that you don't know about, and it's up to the application to figure out hey, what do you do with this data? Basically column? Family databases are aware of what that data actually is in terms of within your value, just like nested key values.

A

If you will so, you can say: hey, okay, I, have this uh book object, but within this book object, I know, there's Fields called of name or ISBN or Price or so on and so forth right. So this is what column. Family databases, One, Step, Above That is a document database, so this is mostly column value, except for the fact that these can be much more complex structures, literary documents.

A

So, if you're, my age, think XML structure, something in some case, something that has structure uh typically again. The the common uh storage mechanism. Here is something like Json, but the idea here is that you should be able to navigate these objects to literally documents, uh so either you can have all kinds of interesting data structures within uh you know your your database effectively and then you have graph databases, graph databases, basically what they do is they are more applicable to things that map naturally to graphs.

A

So social media networks would be an example uh or something like that right, where the, where the relationship between objects is significant and there's actually values stored in the relationships themselves, and there's not that many of these they're relatively there's only a few well-knowns um one well-known uh ones out there, neo4j being probably the most well known one. So aside from that Jacob Donald SQL also has built-in capability of bean validation. Just like in jpa. You can use mean valuation with it in Jakarta SQL you can use V invaluation with it.

A

uh It also it also it's it's. A configuration mechanism is more modern. You don't really configure Jakarta SQL using XML. You actually just configure it using micro profile, config, so models like key value, pair uh property files and that's what you use to configure Thunder SQL, and this is really the reason why it's not. Finally, it because right now, uh Jakarta single has a dependency on micro profile configuration which of course, is not part of the Jakarta platform. But hopefully once uh Jakarta can config goes final.

A

It can use the config as a dependency and then become a final spec. So that's the holdup.

B

A

uh The person, mostly that's behind chapter number SQL, is uh utavia Santana, interestingly he's not speaking at uh this particular Jakarta one but I guess I'm a little bit proxy for him. uh So uh you know, I did talk to him. I asked him hey. What's next uh for Jakarta no SQL, um and basically he gave me a lonely, long laundry list of things, but these are the main things that we'll be coming soon. We're talking to the SQL, uh obviously getting the stable release out. Another interesting thing is Java record support.

A

Right now, jacket on a sequel is based on pojos, uh technically I think it could very easily support uh Records to some extent as well, to the extent that you actually can- um and uh hopefully you've heard about Jakarta data right. So this is basically a Jakarta standard that looks a lot like again spring data jpa and Delta Spike data. So once Jakarta data comes about uh the repository part of the jakartano SQL is really redundant, no longer necessary, because a lot of those abstractions are so high level.

A

They could apply to jpa as well. So, potentially the repository part of jacket that SQL could be defectored and moved into Jakarta data, so this is probably another reason why it may be in a little while before jakartano SQL is actually ready to have a final release. Okay moving forward, um so the implementation for Jacob is called jnosql. There's only one implementation so far uh and uh I personally think it's a really impressive one.

A

um You know looking at all of the different things that are similar, Solutions, um actually Jakarta, SQL and Jane and SQL in my opinion, supports the most amount of databases right. So this is just a small subset. You can I'll uh at the very end of the talk. I'll leave you with. uh Basically what uh list you can go on and check yourself in in the Jakarta single website. The complete list is there, but really honestly, very good coverage right.

A

So this is I, think a strong point for jacket and SQL and as you'll see in a moment.

B

A

Really easy to switch between each of these databases, so you don't even have to make up your mind what kind of noticeable database you would like to use. You could search from one to the other without um too much trouble, thanks to the way that jacket structured but yeah.

A

So this is just a a laundry list of sort of you know all of the different, uh more SQL databases that JavaScript or nosql supports, obviously Osmos baby being one of them uh and which is a segue into of course, me talking about what Jakarta, uh what what Azure Cosmos DB actually is all right. So actually let me uh before I jump into the next section. Let me stop for a moment and ask any questions so far.

A

Oh no I, don't know if you're still uh able to uh chime in or not I, just uh primarily I'm asking you actually you're there there's any any questions. I need to answer at this point.

B

Okay, Iris I'm here uh but I, don't see any questions uh for now.

A

All right reminder, folks, don't hold back until the end. If you have questions just ask them all right, moving forward, um no actually I'm, uh you know I I, don't know I, don't want to say I'm a nosql skeptic, uh you know I would say I'm in general or relatively conservative technologist. If that thing, even if such a thing even makes sense so.

B

For me, it's always a question.

A

Of like okay, what is the use case for this right, um and hopefully uh you know that in about 10 year, in 10-year time frame or around the 2010 uh time frame, there was a lot of hoopla about how like these nosql databases are just gonna, make relational databases Obsolete, and this is the the way to go and why wouldn't every single application use, nosql and so on and so forth. uh So 10 years later, I think we need to take a step back and really think about.

A

Okay, is that actually true, uh and if it's not true, where do we actually use this nosql technology and what are the? What is the data behind it um and, in essence this is my day job also right now, I mean alternately when I make product decisions. It's a lot of. It is driven by customer input and customer data. So it's natural for me to take a look at data and also share some of this data with you. So let's try to understand.

B

A

Does this nosql landscape actually stand today right? So uh here's one way I'm looking at it so I put it put this into Google Trends I also did an indeed.com search. uh Indie.Com is a job site. As you know, it's a it's a sort of a conglomerated job site like it's, it's a job site of job sites, so you can easily look at Job Trends there as well, but actually the data is very similar between Google, Trends and uh and Andy indeed.com. So you can see here what the trend here actually is.

A

So there was a point in time where uh you know this. Nosql databases were actually growing fairly rapidly in popularity, I think that was the early days right so but you begin to see a Slowdown right. uh The Slowdown comes in I, think around 2015 2016.

A

uh and more or less in after 2018, or so in the past four years for five years. It's actually relatively stagnant right, so it has. It isn't really like. Basically, it was subject to Hype in the early days. The hype has basically tied down at this point at this point right. So there's not a lot of hype around us around this.

A

The other thing to notice is that there's very clear leaders and followers in this uh nosql landscape as well. So there's literally dozens of you, know SQL databases, but when.

B

You really look at.

A

The data- it's really just one database more or less that actually dominates uh in the nosql landscape, and that is mongodb right. So mongodb is by far the peer leader, as you can see from this data redis, which is like an old guard. You know key value, caching technology, surprisingly, that is basically like sort of the second best right and it's not even a nosql technology per se.

A

It's more of a caching technology than uh a uh you know, simple technology that comes in in second place, very clear, second place, and then there's everybody else right, so I didn't even put in all of the dozens of different nosql databases like they literally would be like at the bottom of the you wouldn't be even be able to see it. It will be like overlapping lines at the very, very bottom. But third place is something like select Cassandra and you can see.

A

Cassandra is dropping off right, so it's actually dropping off the drop started, something around the 2015 2016 time frame and same with all the rest of them right. So they're at is linear, maybe actually listening in in popularity. If you will but more or less a established landscape, same thing, for uh if you look at all of this and then there's a very interesting website out there, it's called DB engine ranking okay. So this is basically what they do.

A

It's like sort of uh Tob for database engines right, so the tracking sort of the usage patterns for nosql databases and again you can see very clear picture zero. What actually is going on so overall um by a good margin? You know the market leaders really are the Old Guard Technologies. So you know oracle DB MySQL our SQL Server postgres.

A

These are the guys that are at the top right and.

B

Buy a good margin, so what this tells you is canonically when you're.

A

Starting an application, you actually should probably just start with a relational database. um That's in in practicality. That is what most people are actually doing um and again, the ones that actually stand out in any meaningful way is redis, and in this particular uh data point actually redis comes out a little bit on top of mongodb, but really that is in mongodb and then there's the rest right and you can see that uh you know other than at the very top.

A

You know, uh barring things like Cassandra and elasticsearch, which you can argue, isn't even a database. Really it's it's a it's a way of uh doing Consolidated logging more than it is about a database, all the rest actually aren't even tracked because their their adoption is so small. It won't even show up in this graph right. So all of the all of the rest of the dozens of uh non-sequential databases, that's really where they are so that leaves should leave you with an interesting question of.

B

A

Should you actually use nosql? So this is an interesting question, but I'll give you my opinion and I didn't really make up a slide for this I'll, just uh sort of left it as a discussion point um as a general recommendation. What I would tell you is? Probably you should start with a relational database when you're developing an application most likely for 80 to 90 percent of applications. That is the right type of database. You want to use and it fits The, Right Use cases for you where you should begin thinking about you know.

A

Sql databases is a few cases right. So one is if your data is really really large right. So relational database is one of the big problems for them. Is they do not scale horizontally, Beyond, one or two, maybe at best three instances, so that only means that you are constrained by CPU resources and you're, also constrained by I o and constrained by storage.

A

So in general, if your database is more along the lines of multiple terabytes or petabytes right, that's where you should really think about. uh You know relational databases because you're sorry, nosql databases, because relational databases just won't be able to handle that type of load.

B

Gigabytes of data.

A

Relational databases, no problem whatsoever right; it will just perform with applying colors no problem at all, but when you have that scale of data, so we're talking about basically Twitter scale or Google scale or something of that nature uh you know, petabytes of data is a lot of data. uh That's when you really need to think about uh think about nosql databases very clearly, if you have geolocation needs, in other words, the same bit of data needs to show up at in real time across the globe.

A

That's another case for nosql databases, because relational neighbors just won't be able to do that for you, but nosql databases because of the way they are developed with distributed computing in mind. Those type of databases will be able to handle that another use case is really if you have very unstructured data right, so your data just doesn't have a deterministic structure. It varies way too much right from one data point to another.

A

That's another use case for nosql, because nosql database is generally will not enforce structure on you right, so the structure can be quite flexible.

A

In my mind, I think those are really the use cases for when you would want to consider a nosql database over relational database, but you know honestly, for the most part, you should probably just stick with a relational databases to begin with um any questions on that.

A

So let me stop there before next I'm going to jump into Cosmos DB, then I'm gonna we're gonna, stop sort of uh the slide part of the presentation and we'll dive in in more into the code and begin looking at code, but before I do that, let me stop for a little bit um yeah any questions comments. Anything you'd like to discuss at this point.

A

B

A

Guess the answer to this now so we'll just keep moving forward. So now, let's talk about Azure Cosmos Daily Now. What is azure Cosmos DB? Precisely uh so obviously you can run people.

B

Are running you.

A

B

Sql database and relational.

A

Databases just fine, uh you know in their in their data centers on premise: what have you, but what Azure cost almost DB is, is that it is a managed database. It's a pass service by the way we are on on Azure. We also have uh past services for relational databases like Azure, SQL, Azure postgres.

A

um You know Azure MySQL, so on, but basically Azure Cosmos JB is our Flagship. Nosql managed database right. So if you, if you want a nosql database- and you don't want to be in the business of infrastructure management, Azure Cosmos DB- is your answer.

B

I actually have a question for you. Sorry, yeah, okay, interrupt you.

A

B

Do you see no secure being included in the future version of Jakarta Yi and if yes,.

A

Service to be included in any profile, um I think something like jagatan. Merciful is best if it's just simply a standalone specification under the Jakarta umbrella. So this is a recognition that it is part of the Jakarta family uh and we should leave it to run times to include them include something like Jacobs equal in a modular fashion. So great example, with the of that is uh Liberty, for example. Right so I would love to see.

A

You know, uh show up as a liberty, module, uh I'd love to see, see the SQL support added in quercus uh love to see it in Helena. You know, but I don't see the need for like forcing implementations to add it by default in any given profile, mostly speaking,.

A

B

A

B

No, that's it. Thank you.

A

Okay, all right so back to address, maybe what okay, okay, so other than administrative, is what actually makes Cosmos DB so special. So what makes Cosmos really special is uh actually not so much that it is a managed database, but what you can do with it when, if you really have the use case- and you really have the need, so it is globally distributed, meaning you can basically take a bit of data and say this data is going to be available.

A

You know at a global scale and Azure Cosmos TV will do that for you without you even having to do very much else other than saying hey for this database. This is how how it is right. I want all of this data to be available everywhere, I close across Geographic distributions in real time.

A

um It is also a multi-model database. I'll talk more about that in what that is. What that means is that it speaks multiple languages, so it has a.

B

A

Language but other than the native language, it can actually speak, so it can act as though it is. You know many different types of database rather than just a particular type of database. um We also focus in causative. We focus a lot on sort of the enterprising features right, so I've heard at the end of the day is more or less honestly an Enterprise Cloud.

A

um So we feature we focus a lot on those Enterprise Readiness features, so things like availability.

B

A

Uptime, you know even in a geographically distributed database performance, so even in a highly distributed environment, we want to guarantee regularly millisecond availability in in terms of your queries and your data uh scalability, both horizontally and horizontally, primarily right. So this is about scaling up and down your instances right, based on what your data volume is. What and what your volume of demand is, and obviously keeping everything secure right for data security is a number one priority.

A

So if you actually want to make sure that this managed database is secure and the security is handled on your behalf, you don't have to worry too much about it. So those are the things that actually makes Cosmos TV sort of special. If you will um and I talked about uh that, this is a uh what's called a multi-model database. I actually did very few of uh there's very few things in uh in in the industry that actually does something similar, but basically at the heart of it.

A

uh There's a core runtime for Azure Cosmos JB, the core runtime uh is basically key value. Pairs are actually not key values, but really the document database under the hood um and then and then that document database uh allows for querying in a language that looks a lot like SQL and that's the core API for Cosmos DB as well, and conveniently this is actually a superset of what you can do with a nosql database.

A

So it takes it's taking like basically the relation the nosql database ecosystem to the Max and say: okay, how how much can I do with it with a with a nosql database and that's what the core API actually is? You can use the core API if you want um another very interesting thing that we have done very recently with Azure Cosmos DB is azure. Cosmos daily can actually behave as though it is a highly distributed. Postgres SQL database right, so it will basically respect post SQL, syntax, uh even in terms of database creation.

A

So that is another very interesting option. If you're going to do that, basically, you have this relational database that, under the hood, is actually run by a low SQL database. So it's an interesting concept. um It's not! If you're going to do that. You'll probably be using you'll, be probably using jpa and jdbc to access Azure Cosmos DB, rather than any, you think, related to Jakarta, no sequel per se. It can also speak mongodb as an API, so this is basically handling the document database, so it will basically conforms to the mongodb protocol.

A

So anything, any sort of API that can talk to mongodb can also talk to Azure Cosmos DB uh same same deal with Cassandra Cassandra is a column family database, and he saw you know all the lists of things that really normally be a customer are more or less so the top players, if you will and again Azure Cosmos, unique and sort of act as though it is Cassandra.

A

Similarly, it has capabilities of acting as a graph database, so the protocol that it follows there is Gremlin. Gremlin is sort of kind of sort of a standard. There's a there's. A project called Apache Tinker pop that actually implements Gremlin uh Gremlin is is implemented by neo4j, also right. So what this means is, it can basically more or less cause more DB can act as though it is neo4j as long as you're using the granular language and then finally, there's also plain or key value.

A

So the key value API that is in use is something called Azure table. Storage. You probably don't know what that is. Basically Azure table storage is a is a general purpose. Azure technology- uh it's uh if you will S3 plus so Azure storage does does have, does have something that behaves us through as a it's a much like as S3, but it also has richer capabilities. So this is basically your key value database. If you want so it can it can, it can speak that dialect as well.

A

So uh you can actually speak to Azure cosmology via a Jakarta SQL, using these various dialects. So you can either talk to it using the mongodb API. You can talk to it using the Cassandra API. You can talk to it using the gremlin.

B

A

Because this is also something that Jacob from the SQL supports as well, so far doesn't have, doesn't speak or uh you know Azure Azure Cosmos TV. So basically you can speak to Cosmos to be using these other protocol mechanisms. Sql.

A

Okay, so without further Ado, let's jump into a little bit of code um before I. Do that I'll stop a little bit, make sure there's no questions specific to Azure Cosmos DB. But this is the where the code reservides. uh You know you can see the URL I'll post a URL in the chat uh in towards the end of the talk, so you can have it, but it's really really just in my personal repo.

A

uh So uh let me stop here for a moment, while I pull up the browser uh and show you what's in the repo. Thank you.

A

Any questions so far.

A

So I also want to log into the Azure portal. So I'll do that, while, uh while I wait for any questions that you might have and I hope, I'm not having some kind of network outage, that will be really interesting.

A

Network outage in the middle of my talk, that's what you need right: okay, uh I'm gonna continue hoping man uh relatively Zoom is not my favorite uh interface for uh screen. Sharing it's dirty, clunky, okay,.

A

All right, finally, here's my sign in.

A

Here's the Azure portal, okay, I'll, show you the Azure portal in a moment. I don't want to uh worry about it too much right now, um I'll show you what's why I'm showing you the Azure portal.

A

Oh I have the wrong tenant, so this is part of the reason why I was the login was taking such a long time. Let me switch tenants. There was my uh work. Real work uh tenant we're moving into my uh sort of community demo tenant. Okay, all right! So here's a here's, our repo, uh so repo. Basically, my intention with this repo is that you can think of it as a sample of how to use your ethanol SQL with Cosmos DB.

A

The idea here is to uh be be able to be self-service, so you can actually go to this repo clone this repo and actually be able to run, go through the demo by yourself and get started by yourself right. So there's.

B

A little bit of setup involved, I've.

A

Already set all this up locally, so you need jdk, 11, plus that's what chocolate thunder SQL requires and that's what I have in this. In this particular case uh you can it's just an amazing project. You can use any ID that you want I happen to be using Eclipse, because that's what I'm most used to uh you do need an Azure subscription.

B

A

Because Azure Cosmos DB is is an Azure for service, so you need to run into an Azure. uh In this particular case. uh You can get a free subscription for one year. It's not a big deal, you're not going to charge anything, um and in fact it goes beyond that. uh Azure Cosmos DB actually has a free tier as well right. So even after the one year you can continue to use it and not actually be charged. Anything if you want. You will need this code. Example you can clone it or download it.

A

uh Then I will go through the steps for you to uh go ahead and create advert Cosmos TV. All the steps are outlined here. I've actually gone ahead and done that it takes about half hour for the initial deployment to happen because of DB. So I didn't want to do that right in the middle of in the middle of my talk. Obviously, I won't have much much of a talk left.

A

If I did that, although I've been experimenting to think about whether I should start the deployment at the very beginning of the session and then say Oh, you know I created the database for you in real time, but in this case I wasn't super adventurous. I actually went ahead and created these steps, and so I have Azure Cosmos DB instance running I'll show you that instance. In a moment, then you need to run the application very simple.

A

All you really do is open up your ID import, the project as a maven as I may even project which I've done um already and basically off you go right. You can basically just run the project.

A

um I haven't I, didn't actually do anything very fancy and I should just have I just actually have it just a unit test uh to show you that I want to keep it very simple stupid. It just uses this world to spin up a CDI instance, and you know run a unit test for you and that's about it. But obviously you can apply the same thing to uh project if you want and the code would be if anything simpler.

A

uh So, okay, so uh to do for those of you who are open source-minded, um you know I'd love to actually have multiple variants shown uh in. In my code example right now: I'm just using mongodb right, so I'm gonna demo, the mongodb way of accessing uh Cosmos DB using jacket, Thanos equal, but I'd love to actually also incorporate Cassandra and Gremlin. So there will be a welcome PR right.

A

So if you're interested in doing this, you know hopefully get in touch with me, and maybe you can help me do that before I finally have the time to do something like that. So let me uh get past the annoying uh Zoom box. So here here here's my uh here's, my Azure portal. So this is the resource Group that I created before it's. Basically a grouping of different Azure.

B

A

uh In here you can see that I just have one thing in this in this group, and that is basically my uh so Cosmos DB instance. If I go in there, I'll say Azure Cosmos DB for mongodb account okay. So this is a cosmos DB instance uh that is acting as as though it is. It is mobility, so I'll show you uh different stuff, as, as I run run the demo uh we'll just leave this open for now. I should only the GitHub repo anymore.

A

So let's now hop into the ID okay, so here's the project. uh It is basically a maven-based project. I'm gonna, uh just to get things started, make sure I, don't mess things up in in uh in the middle I'm, just gonna say made them clean.

A

Okay and next thing. I'm going to do is I'm going to go ahead and do a run as Maven install and what this will do is it will actually run a bunch of unit tests. So let's take a look at it. Take a look at what happens.

A

Okay, so there are my tests. Bootstrapping.

A

And it went by Superfast, okay, but there's a whole bunch of stuff that happened, but let's go through it, real, quick, okay. uh So uh hopefully you can uh take a look at this. It says: testing persist of orders and customers. It says I saved three orders uh and then testing fine by ID and found in order right. This is how the how the order looks like uh testing querying right, so there's different queries.

B

A

I'm running right now so I'm trying to found orders with cost greater than 1001 of those found orders with a machine too found a testing update order. So here's what what I've updated in terms of the order and then finally I'm deleting the order deleted order with three and also testing Bean violations. I'm testing that uh testing probability of an order and in this case a invalid offer invalid order that I tried to persist and I'm getting a nice error message because of it. So.

B

Let's take a look at the code, real.

A

Quick um I know I, don't have a whole lot of time. I probably have like plus or minus five minutes. So let me see what I can do and work around again my pesky little uh the best little uh thing for zoom.

A

Okay, so I'll show you that the uh the test, first, okay, so here's the test.

A

All right and can I full screen. This can I full screen this. Yes, okay, so uh I have several tests here. uh What I do here is I'm initializing on the SC container for CDI, uh getting it what's called the template. In this case, it's going to be a document template because I'm using mongodb and it's a document type then I'm deleting uh everything from the order table or the order document.

A

uh Here's how I'm creating a customer. Okay, uh here's where I'm creating an order, I'm setting up the relationships, setting up an address for the order, setting up the order lines right, as you can see several art lines and finally, I have my magical insert statement very similar to jpa uh and and adding another one in the same way and adding yet another one and then finally asserting that I have three nuggets: okay.

A

uh Similarly, here's my point by ID so I'm, finding by a given order, ID and asserting that I found it uh I'm doing several queries and because I am using mongodb. I can use syntax like this right, so I can say from order where total cost is greater than one thousand. Okay and I can find out uh whether that that is true or not. I can do a like query to see.

A

Hey uh I want everything that has a pinball or everything that has a machine, so it can support these type of query languages as long as the underlying no SQL database actually supports queries. Okay, I'm doing an update here and then I'm also doing a delete here and here's. My validity I just created an empty order, but when I create an empty order and I try to insert it it's going to fail right because it's an empty order, it's invalid. So if you look at the order object, it is.

A

It uses uh uh a lot of.

B

Annotations very similar.

A

To uh jpa, but not quite the same okay, so uh and I'm also doing Bean resolution constraint here, so it is an order, is an entity. It.

B

A

Id, just like you would in jpa, it has columns. So in my case it has simple columns like description and total cost. All of them, as you can see, have been validation annotations attached to them, but then I can also have complex objects. Okay, so I can have embedded objects. That is the order line. I can have customers and addresses. So a customer is an.

B

A

An address is actually an embedded, okay and embeddable, so very similar, very similar to jpa right and I'll. Just briefly show you, the configuration before I show you how that data actually looks like. So uh if you look in Source main resources, this is how you configure the thing. It's super simple, okay, you simply say you know: what is your data.

B

Store name what.

A

Is the database name, and this is like your mongodb url? Okay, it's a long URL right. Then this is how you connect to mongodb. Just like a jdbc URL and then you say hey what kind of Provider do you want right? So in this case I'm saying I want a communication driver. This is your jdbc driver configuration if you will for mongodb using the the document model, so I'm, just gonna uh I'm, almost out of time, I'm gonna hop back into the Azure portal and show you how the data looks like okay.

A

B

A

Really like about this uh is that it's a relatively self-contained UI and I hope. My network is not gonna uh be dying here. So uh if you look at the order- and you look at the look at the table and look at number one okay, this is how your data actually looks, like my apologies. I should have shut off my my phone, so this is how the data looks like, as you can see it is in. It is in Json format,.

B

A

So this is uh this: is the structure of this data all right? So that's about it in terms of uh let's hop back right into the into my session. Okay I can have another 10 minutes. Okay sounds good to me all right, so, uh let's hop back into into uh into my session all right. So.

B

Record for no SQL.

A

Is basically a new standard like, as you can see, it's a new specification? uh You know it is a as far as an unsql databases go go and manage one goes, it's a very good one uh Cosmos DB, uh and then you can see how easy it actually is to use these two technologies together. So in my opinion, SQL is probably the one of the better apis for for accessing the SQL database. So I'm very glad.

A

It's part of the part of the Jakarta um Jakarta umbrella uh and I hope that the demo that I showed you is is so starting point for you to take a begin to take a look at these things. uh So here I'll leave you with some resources, so the resources here is uh you can, uh if you want to learn more about Cosmos VB, the best place to do that is actually the cosmos DB website. I actually learned want to learn more about Cosmos DB in in Greater technical detail.

A

The best ways to look at is actually Cosmos DB Docs. uh If you want to learn Jakarta nosql the best place to do that is a Jakarta SQL spec itself. That is online right. This is simply janosql.org spec, it's as simple as that and then. Finally, if you want to learn more about uh jno SQL itself- and it actually has some nice starters in the home page and also you know things like hey what what uh what kind of databases does does this thing support?

A

All of that is uh in in the home page as well, so.

B

That's about it.

A

In terms of uh you know what I have I'll post in my chat here, uh I will post the URLs for, uh for my or my slide deck and my GitHub repo in the meanwhile, if you have any questions, let me know and I'll uh I'll answer them: here's the URL for the GitHub repo, and you also give you the URLs for my talk.

A

Okay, here it is here's the slide, deck and I believe there's actually recordings of this talk already out there on YouTube somewhere, but obviously this talk that I'm delivering today will also be recorded in some some shape or fashion.

B

Yeah; okay, with that.

A

I'm going to stop sharing my screen.

B

Thanks so much yes, we're going to share this uh on our YouTube channel and it's going to be on the Jakarta one live stream uh website as well later on, but yeah, as you mentioned, it's being streamed on uh Twitter and YouTube and Facebook as well as well. uh And it's it's going to be available there for people who are interested to re-watch it great.

A

Okay, well, um you know, let me know if you have any last-minute questions, otherwise uh yeah, it's been a pleasure, as always, I think this is like my fourth, it's the fourth uh Jakarta one, and it's my first time speaking here so.

B

Yeah, that's it's been a pleasure. Thank you. So much I don't see any questions, but if there are more questions for Reza, please feel free to add in the chat section and he can reply to you there as well. Yes, otherwise.

A

Yeah also, my contact information is like on my.

B

Website, let me post that.

A

Also, uh uh you can always uh just reach out to me, uh and you know later on and I can I can chat with you. One of the reasons I do. All this is uh is to actually connect with ebook. So there's there's uh I also posted in the chat. My contact contact details.

A

All right well have a good evening. Folks, I know it's been a long day for some of you uh in.