MySQL Database - Full Course

MySQL Database - Full Course

SUBTITLE'S INFO:

Language: English

Type: Human

Number of phrases: 982

Number of words: 16466

Number of symbols: 74308

DOWNLOAD SUBTITLES:

DOWNLOAD AUDIO AND VIDEO:

SUBTITLES:

Subtitles prepared by human
00:00
MySQL is one of the most popular databases, learn how to use it in this course from a senior database engineer. Welcome to this basic MySQL course I want to start by appreciating you for trying to learn a new skill. Let me introduce myself. My name is Barbara and I work for Salesforce as a senior database engineer, I have over 12 years of experience with a variety of databases Oracle being the main one, I have experience working with companies like Chase, PayPal, Wells, Fargo, StubHub, etc. Let me answer a few basic questions for you first, and that is who, what and why. So who should take this course this course is meant for database professionals who want to expand their skill set. If you are a software engineer or a full stack developer and you want to gain a deep understanding of MySQL database, this course is for you. And if you're a college student, or a computer science student or a fresh graduate, this course will give you some knowledge on database internals.
01:02
So why should you learn MySQL, MySQL is the most popular open source database and of course, Postgres. SQL is definitely up there. As companies move their data from on prem to cloud, they usually like migrate to a cloud native database, or an open source database, like MySQL or Postgres sequel, in order to save costs. So let's say that you are an Oracle Database expert. If you gain knowledge on a database like MySQL, then you could help companies migrate their data from Oracle to MySQL, and that can be really valuable. Now let's look at what is being covered in this course. Now, first of all, by MySQL, I mean, the MySQL InnoDB storage engine throughout this course, which is used behind any commerce website, or a bank or a financial institution, and so on. And MySQL offers a variety of storage engines, my I Sam, in memory storage engine, or some popular storage engines, which are
02:04
available, we are going to learn about MySQL InnoDB, I am not covering any other type of storage engines. Now, these are the topics that I'm going to cover in this course. And please note that this is a database administration course. So it is 80% database administration. And for people who are completely new to databases, I have included SQL basics. So you'll be learning about database installation, MySQL Workbench, database indexes, database logs, you'll also learn a little bit of performance tuning, that is SQL explain. So these are some interesting topics that I'm going to cover. So what do you exactly need to get started with this course, you need a PC or a Mac. So if you're using a PC, then I recommend you to actually look into the worksheets or supplemental material that is attached in the description. If you have a Mac laptop, then you are in the best position to learn this course. Because then you can just see what I'm typing. And you can just type the same commands and just follow
03:08
along from end to end. And most of all, this is the main requirement, I want you to create an AWS account, that's correct and amaze on Web Services account. So if you don't know what I'm talking about, please do look into my worksheet which can be found in the description, I've attached some resources, which will show you how to create an AWS account, I will be using an AWS EC two instance throughout the course. And I'll show you how to create one. But one main thing that I want you to remember is after each study session of yours, you can shut down your EC two instance. That way you don't have to pay any unnecessary cost. And please do remember that you don't have to keep your EC two instance running 24 seven. So once you have created an AWS account, and sign in, you will land on this dashboard, or this page. And you can go to the services menu right here. And then under compute, you
04:13
can choose easy to so over here on the left hand side, you can choose instances. And then here we're going to create an instance which will be our lab environment. So click on launch instance. And then let's choose an image for our instance. So I'm going to choose Red Hat Enterprise Linux version 864 bit and my instance type is going to be T two micro which is free tier eligible. And you need to choose an appropriate subnet. If you just created an AWS account, you can just leave. Leave it whatever default subnet that shows up for you. For me, I'm going to choose me Be a specific one. And then make sure to enable this auto assign public IP option. Because that will assign a public IP to your instance using which you can SSH into your instance, from your laptop, and leave all the other settings as is. And
05:24
let's allocate like 25 gigs, for the database, or for the entire instance. And you can just leave the rest ISIS, go to the next page read, you can create a tag for your instance. So I'm gonna call it my SQL instance. One, go to the security group page. And then here, what's important is that you need to be able to SSH into the instance. And you need to, you need to create proper firewall rules for that anyone, basically, anyone in the world can log into this instance, via Port 22. And that's not secure at all. And I will take care of this instance. Because once I finished recording, I usually remove it. So I know
06:27
how to handle this. But when you create rules, make sure to put your laptop's IP in there, just so it's more secure, then now you can review your configuration, and then click on launch. But if you want, you can create a new key pair, and then just give it some name. And then download it before you create the instance. For me, I'm just going to choose an existing key pair. Maybe this one, and I say knowledge, maybe a different one. All right, this one. And launch instance. Now your instance is getting created, it probably take a couple of minutes to create this instance. Okay, that's how you create an EC two instance. And now that my instance is up and running, and I can see the public IP. Later, I will
07:36
be installing MySQL on this instance. So that's what I wanted to show you in this lesson. So what you're seeing is basically my sequel documentation showing all these different installation guides, like Windows and Mac, and so on. So we are interested in Linux installation are basically MySQL installation on Linux. And there's a couple of guides actually. So this one is basically installing jannettek binary, we're gonna skip that and go here. And even within installing MySQL, on Linux, there's a bunch of guides. So the recommended method of installation is using RPM packages from Oracle. But we're going to use this MySQL yum repository based installation. And it's actually quite straightforward. So for this
08:42
installation, we need to go to MySQL comm and downloads. And we are downloading the Community Edition and go to yum repository. As you know, the instance that we created has Archie l eight, Red Hat, Enterprise Linux eight running on it. And so we need to download this rpm. But then we need to download the RPM on the instance itself, the one that we created. So let's actually log into the instance. So I'm going to use SSH, and we're going to use my private key. And logging is easy to user choose the default user and basically get the public IP of my instance login and we switch to a root okay. So one thing that
09:53
we need for downloading this RPM to this Linux instance is W get package. So let's go ahead and install that first. Alright, so now that w gate is installed, we need to download the RPM that we just saw. So to get the link of this RPM, we need to go into this download. And we have to right click here and copy the link. And if you're installing on a different OSS, you need to click on the appropriate button. Okay, so we got the link, and let's just go ahead and paste that link over here, like w get and the link. And that command downloads this package, now we are going to use an RPM command to install this package.
10:56
So this package, as I mentioned before, is going to add this MySQL yum repo to your local system repo list. With Red Hat Enterprise Linux installation, you get like a MySQL module by default. So let's disable that one. If you don't disable it, then this will interfere with our MySQL installation. So let's go ahead and disable it using this command. And don't worry about writing these commands, I will put a link to my Git repo with all these comments in the description. So all these have been disabled. Now let's go ahead and install MySQL community Server Edition using yum install MySQL community server. And let's put minus
11:56
y in there just to go ahead and accept all the prompts. And that's installing all these packages. Alright, so my SQL has been installed. Let's go ahead and start the MySQL database using system CTL. Command. And let's check the status. So now, my SQL database is up and running. Okay, so the log file of this MySQL database software is under var log. And then if you grip temp from this log file, you will get the temporary password for root user. And you can use that to log into the MySQL database. And how do you log in, you
12:57
use this command my sequel minus u, that's going to be root and minus P is for password based login. And then we're logging into MySQL database. So let's use this password and see if it logs in. And we are in. And if you run any command at this point, my sequel is going to ask you to reset the password using alter user statement. We can do this in a different way. So there is a executable called MySQL admin. And this is the command for it. MySQL admin, minus you the user name and minus p password. We are going to reset the password of root user. And let's provide the current password which is this temporary password
13:57
first. And let's provide the new password No. All right, the password has been accepted. Now let me try to log in with this new password using the previous command my SQL minus u proved and minus p MySQL. Let me put the password I said just now we are in. So let's go ahead and run a simple show databases command which shows all the default databases that come as part of the installation. So one more thing that we need to do to complete the installation is to load a timezone file or timezone table as shown here. So if I do a select star, which is basically a SQL query to read from this table, you can see that the table is empty right now. So let's exit out and run another command. To load timezone related data. So
15:05
this is the command. And let's go ahead and run it. And I'm going to go ahead and put my password, and that loads a bunch of data, you can ignore all these warnings. Let's go back to our MySQL database. So if you do as select star, from MySQL dot timezone, again, it shows a bunch of data. So now, you're good. And that completes the MySQL database installation. Alright, folks, in this section, we are going to talk about data modeling. Okay, so database design, data modeling, schema design, these are all interchangeable words, terms for database design is an ongoing process. So you come up with a basic design, when you sort of like create your application. And then as the application, you know, gets added new features, enhancements,
16:09
improvements, you basically iterate over this design, right, you keep adding new things to your design, and so on. So the first thing that you do when you're doing database design or data modeling is understand business data. And then once you understand business data, you have to come up with logical design of your database. What do I mean by that? Well, basically, you have to design your tables, the columns, that goes into those tables, indexes, constraints, like primary key constraint, unique key constraint, not null constraints, default values, foreign keys, these are all various things that you need to create. When you when you come up with a logical design of your schema, once you actually have this basic table design or schema design, then you can look for data redundancy, that is
17:13
basically you see where your data is repetative. And then you start eliminating it by normalizing your tables actually. And that's because data redundancy causes data anomalies. What I mean by that is, when you have like multiple occurrences of the same data, when you're, let's say, updating certain data, you have to update in many places. And if you forget to update even one place, now you you have two versions of the same data in your database. And that creates like data anomaly data inconsistency is along the same lines actually. And all that happens because of data redundancy. So what we are looking at is basically a spreadsheet and the spreadsheet is basically a huge table a large table, right? And, and what what we are going to do is basically design a table for a ecommerce website, an e commerce website
18:15
is basically like an amazing zone, or E Bay, or whatever, like online business, like online e commerce website, like Alibaba, or whatever, right? Let's say like, you have only one table in this database, right? And you start, as the orders come in, through this website, you start putting data into this table, right? You have, you know, let's look at some of the things that you will record in this table, right, you obviously need like a count of your orders. So you might like numbering your orders, and then how it is coming like, you know, is it desktop or mobile? Or what product is it? Like, you know, here I have got like a couple of books, the book titles, and then the price of the products. And then who is the customer customers details, payment details, delivery details, and so on. So these are all part of like, e commerce word, basically. Right? And you have one giant table. And if
19:19
you look at the data here, right? So, you know, here, I've got like a couple of customers buying you know, sort of like two different product products, right. So, and you can see that the data has been repetitive. What I mean by that is like every time I buy the same product, I have to repeat this data, like the first order that came in was through the web desktop website that was brought by birth and then you can see all the details of this customer and all that details about the product and the payment, payment details as well. Then the second order that came was from a different person, but then it was, you know, the order was for the same product and you had to repeat the product information. Right. The third was from the previous customer. But then this time he bought bought a different
20:24
product, his information, the customer information has repeated actually. So there is a lot of data redundancy data has been repetative. So, this is basically a denormalized database, where you have only one table or a handful of tables, we back all the information from your website, or your business into these few tables. Actually, this is a denormalized version of your database, let's actually go look at what else you could do. So what you can do is basically, you can start with this basic denormalize table, and then you can start taking out all the redundant information out of your database or your table action, it first thing I did was I took out like the customer information, they put it in a separate table, right. And I've got only customer details here. And I started like putting a ID number
21:26
for each customer ID or customer number, whatever you call it. Once I take out the customer information, I have the Orders table, the initial table, I'm calling Orders table looking like this now, right? And you can see that I've got like customer ID column over here. And what is this customer ID column? Your guesses? Right? So this customer ID column is the same as what you see over here. Right? So and why do I have that because I need a way to relate these rows. As you can tell, like, you know, these are columns, these are rows, these rows, I need to be able to relate to a customer, right? If I take out customer information, then how can I relate? You know, this table and that table? It's through a common column, or a bunch of columns. Actually, in this case, it's just one column. So customer ID, right? I'm just like putting the ID number over here. And what else actually we can take
22:31
out of this table. So this is one level of normalization. Right? So let's keep normalizing which is like take out the product details. Right? So product details are also repeating. So here, you don't feel that much pain, because there's only three records in this table. What if the table has million records, right? This is why we need to normalize the table. Now you actually take out the product information and move it to a different table. And then I have a product ID column just to number, like ID the products actually, and your Orders table will look like this, then you take out the payment information to a different table. And your Orders table will then look like this. This is basically the process of going from a denormalized schema or a database to a normalized database. When you have your data in a single table, then you don't need to do any joints. So you might ask actually, what are joints actually, when you're running queries, like using SQL, SQL is a language
23:34
right? A Structured Query Language, when you're running commands in your database, you, you can get all your data from this one table if your table database is totally denormalized. Whereas if you have many tables, you need to sort of combine or join the stables and then you have to get the data out. So that is called joining the tables. So when you have a denormalize database, you don't have to do many joints. And that is kind of good in a way because your database doesn't have to think that much to get the data. You say I want this data and this data is available in this table. So it's just very straightforward. Whereas like in a in a normalized database, when you join many tables, then your database engine with which they call optimizer in, in Oracle or in most of the tape most of the databases. So this database engine has to think more as then like, okay, which table
24:34
Should I scan first? And how should I filter the data in this table. And then Okay, I take the result set from that table and have to join with these other tables. So there is so much more thinking there's so much more processing that has to happen on whichever server this database is running on. Right. And because of that, the performance will be kind of flexible. Lower right and it will consume a lot of resources and you have all that happening at scale as in like many operations are happening at the same time, then you basically have slow performance actually, or at least lower than what it would have been in a denormalize database, but at the same time, we are removing so much data repetition or data redundancy is very low, because of that, the storage needed in a normalized database is much lower. So, you cannot actually like generalize and say a normalized database will always be slow
25:39
or denormalize database will be fast, it is all like it depends actually you have to look at the data and see how much repetition is happening, etc, etc. So, but generally this is how it goes, as you go through this design process, right, you know, see what we have done actually, right. So, we have decided, the tables that we need, like you know, we have orders, table products got the customers and payments, and we decide the column names. And then not just that for each table, you need to decide actually what will be the primary key. What I mean by that as a primary key is a unique key and which cannot be null, actually, which is very important. So using this primary key, you should be able to identify any record in this table, any row in this table. For example, if I say here, the primary key is order number, then I can, anytime if I have an order number, then I can look up
26:43
this table, let's say order number equals two, I can just get this record out of my database. And then you need to also have like some unique keys actually, right. So unique keys is pretty much like primary key. And a unique key can be now a primary key cannot be null, as I mentioned. And then you can also have indexes on your table. So indexes are ways to basically select your day data faster. Let's say that I often search this table based on a customer's email, then I need an index on customer email column, right, you need to decide that. And you need to decide about which columns can be null. Right? Here, none of the columns can be null. Let's say you have another column called preference a the customer preference as in what kind of shipping or what kind of or which phone number is preferred, or something like that. So that can be a null column, right? So you
27:47
can have none columns. Otherwise, you define your columns as not null. Let's say like, in your Orders table, you have this delivered column, when an order is basically created when a customer buys a product on your website. Of course, it's not delivered immediately, at the time of order creation, the delivered column will always have no or n, a n value, right? All these things, all these decisions that we are making we're talking about is part of schema design. And once you have all this figured out, you can put the information in your design tool Entity Relationship design tool. And in the next section, I will show you how I do that on sequel workbench, my sequel world workbench, basically, you can actually then have a pictorial representation of your logical design of your database, right. And that's basically what you call an ER diagram. And of course, you can talk about the relationship
28:52
between the two tables, let's say you can say oh, this table in this table, they have one to many relationship, for example, each customer can place many orders. So that is actually a one to many relationship, right. But one, you know, one order can be done by only one customer. Right? So, so that kind of thing. So you have one to one relationship, one to many relationship, or many to many relationship between tables. Actually, these are all part of data modeling. But you don't need to be worried so much about that, as long as you have a clear idea of like what data is coming into your database. And along the way, you need to define like the data type of your columns. Actually, that is very important. Your names are going to be a wild char. You know, phone numbers can be numbers, and then email is again like a watch chart. And your ID column or number columns are going
29:55
to be int or number. These are all some decisions that you would make you In a data modeling task, actually, that's pretty much what I want to say about data modeling, then there's much more we can talk about it. And like atomicity, like you, you have all the address, sort of like attributes packed into one column, we, we need to split that as well. So that is called atomicity. You can have address separately, city separately states separately, and, you know, zip code separately, right. So those kinds of things, there are nuances that make your database more and more efficient. And, of course, we're not going to go into a lot of details there. But this is the basic data modeling that you need to understand. And as I said before, in the next section, I will show you how to take this and then input that on my sequel workbench. quick recap of what I did in the last section, I basically created
31:04
a logical design of an e commerce website. So what you're looking at is a table that I started with. It's a denormalize table, and we basically took this denormalized table, and we normalized it. As you can see, there is four versions of this table, I'm calling this table Orders table. So there is four different versions. And I with each iteration, I took out repeating data. So finally, we landed with four tables, apart from the original Orders table. So now we have also customers products and payment. In this video, I'm going to take this now all the stables and then I'm going to take the structure, and I am going to create a logical design. Okay, so let's actually go to MySQL Workbench and I am already connected to a database, what I am going to do is go to File and go to new
32:05
models. so here we can add a new ER diagram, an entity relationship diagram. And let's call this database eecom. Store, something like that. So let's go ahead and start creating our tables. Now, I'm not going to be creating all four tables that would take probably more time or long, and I'm going to be creating a couple of tables. And that should be enough for you to understand how we're doing this. So let's just start with the Customers table. So the customer table has five columns, this is the icon for creating a new table, you can drag and drop, or you can try to draw now double click and then create a table called customer and here we can start putting the table the column names customer ID, and then this is going to be populated by your sequence. So sequence is a database object. And it's
33:09
going to be an integer. So we can leave it as is. And we can have it as a primary key, that's fine and a primary key has to be populated, it cannot be now. So that is automatically selected. The next one is customer name, we could split that into first name, and then we can choose watch our and maybe give a little bit more room as in the length of the name and then last name, I can watch 100 and then all these cannot be now so we can choose not null constraint. So these are different constraints which are available. Let's go ahead with the next one address again address. And if you remember I talked about atomicity. So you want your columns to be you know, atomic in, in the sense that here, basically the whole address is packed into one column. good practice to actually split that into atomic columns
34:15
as an address separately, city separately state separately and then zip code separately. So we have all these, of course, none of these can be now and what else is there? So customer phone number, phone number is going to be all numbers. But then I want to make get 10 numbers, of course not null and customer email. So I can just say email id 100 Okay, not null. So since I Id is the primary key here or customer ID, I want to make sure that that we have a constraint to avoid repeating customer information. For example, if you have one customers data for ID one, I don't want the same customer, customers data to repeat for a different ID,
35:21
for example, Id two. So I'm gonna actually make email id unique for each record over here. And then may or may be phone number also. So these are all unique key constraints, or unique constraints. That's it. So we have the Customers table created. So let's go back and see what else we have. So let's now create, I would say product. And then you basically do the same thing, select that is for creating a new table. And then now here, you can just draw in this one, I'm going to call it product. And we want to go through the same process and then put the product column names in there. If you're wondering, this is the same customer
36:37
ID column that we added over here. And we are going to make that a foreign key in a minute. So let's go ahead and split that into multiple columns. Because, again, everything is packed into one column, which is not a good practice. So let's say let's call it credit card number. If the customer is using PayPal, then we need that email. So can use email id over here. So this can be null or not null based on what payment type is being used. So that's okay, so expiration date is going to be a date column. So let's actually change that. So if you're not sure, you can hit that the drop down and then choose your
37:45
proper data type for each. The other thing that I mentioned, which is basically about foreign key, this customer ID is the same as what we added over here. So let's actually make that customer ID a foreign key. So we can just call it customer ID foreign key one. And then the table that is going to be referenced is the customers. And the column is going to be customer ID. And that's it. So you can see that now we have connection or a relation between these two tables, I'm going to actually just add the Orders table as well. I've created
39:02
the Orders table as well, which is the main table and I'm going to now create some foreign keys for the orders. Everything is done. If you want to create any indexes at this point, you can do that. So I guess we are done. So we are basically added four tables to our logical design these four tables and then we have created columns and then define their data types. And also we created the foreign keys and of course primary key and unique key for each of the tables. And you can see the foreign key relationship you know clearly showing here and that You know, that's what you would do to create a data model. Alright,
40:04
so now actually, let's just go ahead and create a SQL script for this data model. So you go to database, and then do forward engineer. And then basically, you provide the database details where you want to create this, these tables are the schema. So this is, these are my details continue, go to the next one, provide the password. Right now we are connected, I had to try the password two, three times. And this has basically created SQL script for us to create the schema and the tables with all the primary key unique key and foreign key constraints. So what we can do is we can just continue and then now the database or
41:09
the schema is created as it goes through and then executes that script. And close. And now you can see the stables are actually created. So you can even go to your SQL editor, and then you can start reading your can start querying your tables. There you go. So you ready came back, of course, there is no data in it. And you can now start using your database. So we actually successfully created the basic schema, or designed the data model for this e commerce website. Table creation or a CREATE TABLE command starts with CREATE TABLE keyword followed by the name of the table and followed by parenthesis. So within the parenthesis,
42:14
this pair of parentheses, you have all these column names, followed by the column data types, and followed by the constraints. And you can also use this auto increment keyword, if you want your column value to be incremented. Automatically, as you load values to the as you load records to this table, and after the column definition, you have the option of specifying the keys like primary keys, unique keys, foreign keys, and so on, you can also specify the storage engine type as part of your table creation. And this is a very simple table. You can also have partition tables or partition tables that have compressed tables, encrypted tables, and all these things require special keywords to be used in your table definition. And please check my sequel documentation. If you want more details about the syntax. As mentioned, I'm using just integer watch char data types, the MySQL documentation
43:20
shows like all these different data types like numeric, date, and time data types, string data types, like the ones I'm using, and JSON spatial. So these are all available in MySQL for you to use. So let's go ahead and create this table. And before creating the table, I want to run this drop command just to make sure the table doesn't exist. And I'm going to be creating that table and see if the table has been created. Yes, the table has been created successfully, I'm going to be running a select star from the table name to see if I can successfully query from this table as well. And then it returns. Basically, it doesn't return anything that means no data exists in the stable. And that's how you create a table using CREATE TABLE syntax. Finally, there is actually a default keyword which helps you specify default values for a certain or for your columns. So if you don't specify
44:21
a value for this quantity column in your insert statements, or when you're loading through procedures, loading data through procedures, it will automatically take this default value. That's pretty much it. And I'll see you guys in my next MySQL SQL session. So I just did a describe on the table that I created and it has product ID product name, product type, price and quantity. And you can see that product ID is also an auto incrementing column right now there is no data in it a typical insert Statement looks like this, let's go to insert into keyword and the table name a bunch of columns within parenthesis, the ones that you want to populate, followed by the values keyword. And followed by the actual column values. If you can realize I haven't actually specified the product ID value because it's an auto incrementing column. So let's go ahead and execute this and insert statement goes
45:24
through. So let me also run the select statement. As you can see, the product ID table has taken the value one, and that's happened automatically ident supply the value one, so I'll go ahead and commit the change. And then let's actually move on to the second variation. So this time, I'm going to specify a value for product ID, nothing else is different. So just want to show you that it is possible. So it goes through and then a commit. And then let's do a select to make sure the value has been inserted. So let's go to this third variation of this insert statement. So it's going to be pretty much the same, except Actually, I'm going to just jump some values and then insert the value 10. For this product ID column, I'll go ahead and do that. It goes through a comment and select again, then you can see that that is also fine. So yeah, so that worked. So you can actually jump a few values. let's
46:25
actually look at the next variation. Again, I'm going to insert a record into the stable, with no product ID specified exclusively. Or explicitly, the product ID column is missing over here, and I'm going to run the insert statement and commit. And then I'm going to run the select statement. So just wanted to show you that wherever the latest value is for this auto incrementing column, I inserted the value 10 for product ID last time, and then the next time I do an insert is auto increment kicks in and then you know increases, increases this value from 10 to 11, right picks up from the value that was inserted last time. And I'm gonna just take another insert statement, and this time, it's actually insert into the table name. And instead of the values keyword, specifying the column names, values, etc, we actually select from a different table. Basically, if the products
47:30
three table is exactly matching the structure of products, one table, then we can do even a select star from if the columns don't match exactly as in like products, one has a different set of columns, and product three has a different set of columns, then we need to make sure that we actually select the columns. And then, for example, this product ID from products, three maps to this product ID and products one, and product name from products, three maps to product one in I'm sorry, Product Name and product one, and so on. Let's go ahead and run this and see what happens. And then that goes through. And then if I now select the products, one table, you can see like take all these rows are inserted properly. And basically the products one table is populated. And then we got all the data from products three table. So this time I want to show you the insert statement. Again, this is kind
48:34
of like a bulk insert or multiple insert just combined into one statement, you can see the insert into clause is specified only once but in the values spark in this clause, actually, we have two rows specified at the same time. So we can even use such as syntax and a commit, and then do a select all good. So these are a few variations of insert statement. I hope you understood how this works. I will see you guys in the next session. Hey, my sequel learners. So in this session, actually, I'm going to be talking about update and delete statements. So as usual, I'm going to be using my eecom store schema. And I will be using my products table to do this demo. So just quickly, if we select products table, there's two rows right now. So the first update is just to show you the syntax of update. So
49:38
you have the update keyword followed by the table name followed by set keyword. And then you can have as many columns as your table contains, but in this case, I've got only one column and I can just run this update. So let's actually add one more clause to it, which is the where clause and this is to just update the rows that Do you really want to update so we will be updating only the rows with product ID equals one. So let's go ahead and do that. And then I'll just come in and do a SELECT FROM products quantity has gone up by 50, it went from 299 to 349. Now one more thing to realize is actually like you can, you know, you can specify literal values, when you are updating, you know, that happens all the time. Or you can also like specify a formula, or you can have SQL functions like replace substring, length, and so on, you know, you can look at my SQL documentation to see what kind of functions are available in this update, like, I just wanted to show you the syntax, basically, to let you know that like you can have multiple, you can update
50:44
multiple rows at once. In this case, I put like, values one, two, and three, you know, you don't have to do one row at a time or anything. So when you use the end keyword, and you can say product ID or whatever, call them in and then a bunch of values to select the rows that you want. And there are other ways to do it. But the point being, you can update multiple records at once. And another interesting usage is using the case statement, you know, you can let's say you have a bunch of update statements, one for product ID equals one another for product ID equals two and another for the other product ID values. And you can combine all that into one UPDATE statement using a case when then and clause or keyword basically, in this case, actually, for product ID equals one, I want to increment the quantity by 50. And product ID equals two, I want to increment the quantity by 100, and so on. So then I've got this similar WHERE clause are similar to the one that I showed you before,
51:45
I'm going to run this you can see the columns are getting incremental, I'm not going to go back and check I'm pretty sure that it's done the right thing. So the next one is basically when you want to delete records from a table or purge data from a table, then you can just use a simple delete statement. And if you are wanting to delete a particular row, again, similar to the update, you can use a where clause to actually like narrow down the data that you want to delete this particular statement, which is delete from a table name, and then where column name equals or the column value. And you can have multiple filters in here. So here I don't have the row number three, I think I deleted it already. Alright, so let's keep going products three table, I just wanted to show you it contains a lot more data than my other table. You can see it contains data about 5849 rows, I wanted to also show you this particular parameter MySQL configuration parameter to basically enable and disable
52:49
safe updates. So let's say like if your delete statement or UPDATE statement is not using a primary key column in the where clause, you know, then basically if you enable this particular parameter, let's say by setting this one, and then if you run your delete, you will get like an error code 1175 it mean, and then it says you're using safe update mode, etc, etc. It's not leading you to the run this kind of add delete statements, because it could be a good cause bad performance. So if I disable the same thing, and then if I run the Delete statement, and then just run a select again, then it should go through because now the parameter is disabled. And two more things. One is actually like if you have a huge table and you want to delete only a few rows at a time, then you can use the limit keyword to limit the number of records that are deleted by the statement, you know, in this case, because I wanted to delete only 10 rows, let's go ahead and do that. And it
53:54
should work just fine. And then if I do a select, you'll see the difference in the row count actually, now it's like 5839, before it was fired four nights. So that's how the LIMIT clause helps you also in the limit floors, you can also specify the ORDER BY clause. It basically sorts the data by these columns, first by quantity, then by product ID, then it deletes the top and or 100 or whatever value you put here, actually. So let's go ahead and do it and then select again. Yeah, Kearney, seven D is gone. So the top 10 rows are gone. And yeah, that's pretty much it. Actually, those are all a few variations of update and delete statements. And of course, there's lots of tangents we can get into but I will leave that task to you. And I hope it was useful. And if you have any questions, let me know in the comments. I'll see you guys in the next session. MySQL learners.
54:56
So in this session, we'll look at SELECT statements, not just the syntax But also like some ways you can actually like improve the performance of your queries, I'm going to be using the schema called income store to explain about this SELECT statement in its most simple form will look like this. So you have the Select and from keywords and then after the Select, you specify the Select list, which is the columns that you want to select. If you specify a star, or asterik, that actually selects all the column columns from this table, and then after the from keyword, you specify the table names where you want to select the data from. So if I do a select star from products underscore three, it's going to return all the data from products underscore three table. But do remember that anytime you are using a star after the Select, or in the Select list, you're basically querying all the columns in this table, you don't need to query all the columns in the table in most of the cases,
55:56
so you only like specify the columns that you need to query. So in this next query, let's go line by line and see what changes that have done to this query to make it better. So let's say I want to select only these columns. That's why I specified only these columns in the Select list. In the from clause, I have specified products underscore three table very often you will be selecting from multiple tables, you need to join the tables and then retrieve useful data out of it. And in the where clause, you specify all the filters, or the conditions based on which your data will be filtered out. So here I am, including only the data which have quantity less than 25. So this way, I'm able to actually filter most of the data out of this table, this is very useful in minimizing the amount of data that you retrieve from the database. And your queries are going to be fast as ordered by is basically going to sort the data that is retrieved based on the columns that we specify
56:57
here. So here, I'm just like ordering by product name. And of course, like when you're sorting data, especially when you're sorting a lot of data, the operation can be expensive, unless your source buffer size, that is actually the memory area where the slots happen. Unless it is sized properly, the operation can be really slow. So you need to pay attention to that configuration as well. And I have this other query, which just goes to show you that like this is a very simple SELECT statement. Again, in this select, actually, I have only the Select keyword and a function, I'm using the now function. But there are several other SQL functions that you can use in the script, for example, I can use the database function to return the database that I am actually connected to. And as you can see, I am actually able to invoke multiple functions in the same query. So that's pretty much it. I'll see you guys in the next session. In this session, I will be teaching about
57:57
SQL joints, let's dive straight into the demo, I will be using a schema called eecom store. And I'm creating a table first called T one with one column, the column name is C one, and I'm inserting these two values in this table. One and two, I'm creating another table called T two with a column called C one and inserting these two values again, into table D two, one and three. So it's one and two here and one and three, here, we're going to go ahead and run a commit to make my changes permanent. So I'm going to be just creating these two tables just to show you the records, D one has one and two, D two as one and three. Of course, a join is an operation that joins two tables. And we have all these different types of joints, we'll go one by one and understand what they are. So this is the syntax, so select an a column list that you're selecting. So we're joining T one and T two. And then we're
59:01
specifying the kind of join that we are making. And then we also have this on keyword. And then comes the condition on which the table is joined. So I'm going to go ahead and run this query and see what happens. As you can see, this query, this inner join has returned the value one, so that means actually, so it returns the values that exist in both the tables that match. So that's what INNER JOIN does. So let's just change it to a left to join, go ahead and run it. Now the left join is gonna return all the values from your left to table which is T one. So T one has values one and two are the rows one and two, and then T two, it's going to return only the matching values. And then for this value, which only exists in table T one, it's going to return and now and then I'm going to change it to right join and as you might have guessed it's going to return all the way Use from table t to in the places where there is no matching value, it's going to return a null.
01:00:06
So let's see if that happens. That's what we expected. So we got all the rows from T two, and then for three, there is no matching value and T one, you know, that position has no value. Now we'll jump quickly to a union. And then we'll come back to a full join, a union is basically going to look like this. So two queries, and then in between, we have the union keyword, let's see what it returns, you can see that it's written one, two, and three. So that's actually the rows from both the tables, but it's kind of like combined the data and then smashes them together. And then you have one, two, and three, and then let's run the same query with a slight difference. We'll put union all and then we'll see what happens. That's written one, two, and one, three. So that's returned all the data from both the tables, but except this time, we have duplicate values, Union gets rid of all the duplicate values, it's almost like a set where you have a unique set of data, a union
01:01:08
all returns all the values, including duplicate data. Jumping back to full join, we don't have a full join keyword. So rather, we do full join this way in my sequel. So basically, you have the similar query where you're joining T one and T two, a left join first on this, there's one column that we have, and then you have another query, again, joining T one and T two on this, just one column, but then we are doing a union of these two, and that's going to return the data from both the tables, we have one, two, these two are matching, then for two, there is no matching value. So it returns and now for three, there is no matching value in T one, it returns a null over here. So this is a full joint, that's basically all the joints, all different types of joints that you can do in MySQL, I hope this example was clear. And I'll see you guys in my next session. All right, my SQL learners.
01:02:09
In this section, we're going to learn about locks. More specifically, I want to talk about the isolation level section. So the first thing is just see what I have here, I have two terminal sessions. One is in black. The other one is in slight maroon color. So I'm actually going to log into the database as the root user. And I'm going to do the same thing over here. There you go. I am logged into my SQL database. So I have a little script here to create a dummy table called T one. Okay, so let me show you the SQL script. At this point, actually, you might not understand the SQL syntax and so on. But then let me explain. Briefly, first thing I'm doing is setting auto commit to zero or commit is basically a command that you use to save your work. Basically, the data changes that you're doing
01:03:15
is permanently stored in the database. When you issue a commit command, in my sequel, you have this variable called auto commit, which is turned on by default, meaning all your commands will be automatically committed. If you don't turn this off, I want to have more control over what I'm doing here. So basically, I am doing an auto commit, disable first, so and then I'm starting a transaction. And just to be safe, I'm dropping this table if I had already created it. So this table doesn't exist. So it says unknown table. And the next thing is I'm creating a table called T one in eecom store schema. And then the column name is C one. And the data type is int and its primary key. So and then I'm actually inserting value, just one row into this table called p one, right, the one that we just
01:04:19
created. And I issue a commit command, alter or the alternative to commit is rollback command. So which basically rolls back rewards the changes that you just done in that session. So if I just do a select star from the stable, then I'm going to see this value, so which is fine so far. So this is pretty straightforward. So far, we haven't talked about the isolation levels. So what I mean by isolation level is when multiple sessions are trying to modify or access the same data data, then you need locking mechanism to make sure the data is not corrupted, or the database is behaving in a way that you expect to see how you actually set isolation levels. And this is the command. So this is the other session I had opened
01:05:20
show session variables like isolation. So that shows like the transaction isolation level is set to read committed. Right. So this is one of the possible options action. So this is read committed, and you have read uncommitted, and you have repetative read, or repeatable read. And then you have a serializable value, actually, so let's go one by one, right. In this session, I already started a transaction. So I'm going to actually try to update this value using an update command. So basically, I am updating the same table, and I'm updating this column to to where the column value is currently one, right. So I'm going to do that the auto commit is turned off. So it's not committed yet for the start a transaction over here. And let me run a query against the same table and just copy and paste the
01:06:23
table name, want to type it, okay, so we see the value one, which is the previous value. And if I ran the same query over here, in this session, I see the value two, because this is the session where we are modifying the data, right? So and I can see the changes before committing in the same session here. Actually, since the value of this transaction isolation, or the isolation level is set to read committed, it is possible only to read the committed data. In other words, when multiple sessions are accessing the same data, in this case, this column right here from this table, apart from the session that is actually modifying the data, the other sessions can only see committed data, any data that is committed just before this select is executed. So I'm going to go here and run a commit, and come
01:07:27
back over here and run a select. So now you see the latest data because that commit happened before I ran this query. Now let's talk about read uncommitted isolation, setting actually freshly log in again, because these things can get tricky. So every time I want to just recreate the tables to remove any confusion. So let's actually log in again, okay, in here, and I'm actually going to execute the same script that I showed you before. So just disabling auto command, starting a transaction, dropping the stable and recreating it, inserting this value, and then running the command. So now here, what we could do is go ahead and update this value to two. But remember, I haven't committed this data yet. Let's go to this session. And here, go ahead and change the setting to the isolation setting to read uncommitted,
01:08:34
because by default, it is always set to read committed action. Right. So you can see that here. So and this is a session level setting. And you can also change it at global level. But for the purpose of this demo, we just need to change it at the session level. So session level isolation initially read committed, then I ran the set session transaction, isolation level, read uncommitted, and then checking the value again. Now it's changed to read uncommitted, if I ran a select star from this table, then I get the value two, and if you remember that I I only updated the value from one to two. And you can already see this beta even though it is not committed over here. So that is how read uncommitted works. So
01:09:36
there is not much locking going on here. Because database is now letting the sessions do dirty reads because one session is able to read and other sessions changes even before the commits happen actually, right. So those are dirty reads. Yes, actually. So let's go on to the next one. So we have seen read, committed and read Committed so far. So now let's move on to repeatable reads action. Right? So exit. So here I am going to just commit. And I'm gonna re rerun my initial script just to clear the table. So drop table and then recreated insert, value one again, and then commit. So now, the table is back to how it looked before. So here, let me log in again. So this is repeatable read setting, right? So remember that the default value for this isolation setting is always read committed. So if I
01:10:45
change it to repeatable read, write, and then check the value again, then you can see this. So and again, remember, or show variables is the command to check the current value. And then set is the command to set the configuration right, so I will put all these commands in like a git GitHub repo file, then you can actually grab the commands from there, and then you can try them yourself. Basically, I'm changing the I'm changing the setting from read committed to repeatable read, right, so I'm going to just start a new transaction over here in here, I'm going to update this value to two. And over here, I'm going to run the Select query that we saw before just selecting everything from this table. And you see that the value is currently one. And that makes sense. So let me go ahead and run
01:11:49
commit. And if I ran the same query, again, I see the value one. And this is the same as the value that was read before, even though the data was changed by this other session. And then committed within this transaction, the data that we are seeing is the same in in other terms, basically, we are, we are reading the same data, or the reads are being repeated. Right. So that is the third setting. And the last one is the most strict locking configuration. So which is called serializable. So I'm going to, as usual, I'm going to drop the table and then just recreate them recreated, inserted value one again, they might come in. So here, we're going to log in again. And as usual, the default setting is read
01:12:54
committed, right? So let's check that first, just to show you, and then I'm going to change it to serial serializable. So what this means is, basically, I'm going to start a transaction. So on the first session, I'm going to run an update, basically changing the value from one to two. And here, I'm going to start a transaction, and I'm going to run a query on that table. Right. And now this query, even though it's just a select, select is just a read, it's not updating, it's not deleting or doing anything, it's just a read, it is waiting, because the update is basically updating this data. And then it's not database, MySQL databases, not even letting this read or the Select query from the other session to see the data. So this is the most strict setting action. So if I do a commit over here, then
01:14:00
on this other session, you will see that the Gradius return and it's seeing the latest value, right. So if I go ahead and run another select, of course, it's returning the same thing. But if I try to update this value from two to three, another update, that is basically going to wait on the Select, basically this transaction that is running right now because the Select again, select is just a read, it is just reading the data. But still it is locking that row in the database, and it's not letting any updates or modifications to that data. And then you can see that the update even failed because it waited for some time and then the timeout value exceeded so we don't have to go into those details. But I'm going to try updating now. And here. I'm just going to exit out of this session, which will
01:15:01
release all the locks. And that will help the update to go through. And then I can commit and exit as well, and how you hope it was clear to you guys. And if you have any questions, please put it in the comments and reach out to me somehow I know you can figure it out, as you guys are next section. Hey, MySQL learners. So welcome back to this new section of my MySQL tutorial. So in this video, or in this section, we're going to talk about locks. So what are these locks? Right? So let's actually approach this kind of like logically. So if you have a database, and if you're the only person working in this database, then you basically need not worry about anything, right? You know what you're doing. So you will insert data, delete, or update data the way you want. And there is no one else trying to intervene
01:16:03
or interrupt your work. But unfortunately, that's not the case. In today's world. If you think of a busy ecommerce database like Amazon, then then there's like, a lot going on on the on those websites. There's like multiple people browsing is like, a lot of people buying stuff. There's the people who are selling stuff on on these websites, they're updating data relevant to their products. So that is basically concurrency, right? So you have many users trying to do something on this website at the same time, so how do you manage this concurrency, that's why we need locks. So if I let everyone work on the same data at the same time, then there's going to be a lot of confusion. And we might end
01:17:04
up losing some data. So let me actually show you a simple example of how that happens. So I have a table, a product table. So if you've been following my tutorial, thus far, we talked about this table called products. So where we store all the product information, right. So now, there's a couple of records over here. And let's say that we have a seller and a buyer who are working on these records, especially like this particular record, the first one, which is a book, and the books, prizes, this and the quantity, the thing we didn't have quantity when we talked about it in my previous sections. But then I added quantity here. So there's this quantity column. And there's a there's a seller and buying
01:18:05
buyer interested in this record, let's look at this, right, so we have sort of like a time sequence here. So what the seller of this particular product is trying to do is he's trying to update the quantity of this product at nine one, he is adding 60 more quantity to that product, which is you know, 40 plus 60, which 100. And that's what we have over here. So then a buyer comes and he looks at the quantity. And then he basically wants to order two or these books, that's 100 minus 90 100 minus two, it's 98 and then you have the quantity 98 over here. So this happened in a sequence. So but we are worried about concurrency, right? concurrency is like when things happen at the same time. But what
01:19:08
if Okay, first the seller comes and then he reads the quantity of this item. Initially it was 40 and then buyer comes and he also sees that the quantity is 40. Right? And at 901. So the first two operations happen at the same time at nine or one seller comes in he says I want to update I want to add 60 more quantity, like meaning I have 60 more books of this title, but then buyer comes and he says okay, I'm buying two items or two of these books. So but while you that he saw before was 40. So 40 minus two is 38. So he updates the quantity 38. So the seller updates at 200. But then, because of this previous look up, the quantity is updated to 38. due to which this whole thing, this
01:20:17
whole operation is lost. And we end up with sort of like corrupted data for this quantity column. So this is a simple example of how concurrency when not managed well might cause issue data issues like this. MySQL learners. So in this video, we are going to look at basically how table locks works. In the context of e commerce database, we created a simple database or schema called the column store. And we created a bunch of tables or used another dummy table to explain our transaction isolation levels. So if you haven't seen my previous material, go back and check it out. And come back here. But then yeah, you have four tables
01:21:19
for main tables. And the main table that we are interested in is products table here. And in the products table, I inserted a couple of records. These are dummy records. So I don't have a front end or application running over here. So we're just looking at database, right. So what what's going to happen in this tutorial is, so we, we're going to basically simulate a situation where a seller is trying to update the quantity of the book that he is selling on this website, which is this first book actually, the common path to uncommon success. And then the right now the quantity of this the quantity available. You know, for this book is 40, right? So he wants to update this quantity 200. And also, we'll
01:22:20
have a couple of more users, or buyers, basically one buyer is trying to buy the same book, we'll have another buyer Hill, who tried to buy a different book, which is this book, tiny habits, and then the same buyer will also try to browse the website, like of course, like, we are going to have to imagine a little bit because I don't have a front end to show you everything. So let's actually see how this goes. So first of all, you know basics first, actually, let's actually turn off the auto commit. Just so just so actually, we have more control over what's happening. And let me do that in all the three sessions I have open and the first session is the seller session. The second session is the buyer one
01:23:21
session. And the third session is the buyer to session, basically. So I'm going to turn off the auto commit, which is basically a mechanism that commits automatically if it's enabled. And I don't want that. So I'm disabling it. So next is I want to show you the transaction isolation level. And we talked about it in my previous session. So right now it's a repeatable read. And it's the same for all. So we are going to change that to read committed, because read committed is isolation is the right isolation level for OLTP databases. So now let's actually start with the first seller session. So three sessions. So the first seller session is going to update the quantity of this book that he's interested in or his selling action. But we
01:24:30
are going to take this aggressive approach and log the whole table. Right. So let's say the application is returned in a way that it logs the whole product stable for right. And then the other session, let's say by one second session, buyer one comes and he is going to try to buy two books and and how actually we're going To do that is by running an update. So we are basically updating the products table and we are subtracting the quantity by two, which means actually the we are buying two books. And which book is there in the book? Where are the record where product ID equals one, right? So if you remember the data, product ID one is this book, let's go ahead and run this update in the second session. And it's going to obviously, wait, because the table itself has been locked for
01:25:34
right by the seller session, the buyer, one session is waiting. And let's go to the buyer to session the buyer to Australia trying to buy a different book, which book is it this other book, which is tiny habits book where product ID equals two. And we're gonna do that. Of course, even that is hanging or waiting. And that is actually a little bit crazy, isn't it. So just sellers trying to update the quantity of this one record with just one book. And everything is tanking. And the buyer, too, was trying to buy a different book, he kind of gives up. So he moves to a different session. And instead of buying or trying to buy a book, he just tries to browse the website, which is a select query or read query, read a select
01:26:41
query, which is also hanging. So the buyer too is getting frustrated right now. So you can see how restricted this kind of sequences. So if someone's using table logs, that's going to basically reduce the concurrency of the operations that can happen in this database. So that's the main point here in this demo. Hey, my sequel learners. So in this session, we are going to take a brief look at row level locks. In my sequel, I have three sessions, I'm already connected to my ecommerce database, MySQL database, and this is how the data looks now. So we have a products table which holds you know, this data, only two books now, just dummy data that I created this, this is the price and you have the quantity column showing
01:27:49
you how many, how much quantity is left for each of these books. So the first session is seller session. The second session is buyer session, we can call this buyer one session. And the third session is a buyer to session. So this is the data. And just for clarity, actually, I wanted to show you the transaction isolation setting, which is read committed. And the auto commit is turned to turned off basically, it's disabled. So unless I commit explicitly, my transactions will not be permanent. So let's actually start with a seller. He's going on the website or a portal that he has available to update the inventory of, let's
01:28:55
say the book one, it or the product one, which is this book. And so he is going to click some buttons, which is going to translate to an update statement being executed in this database, right? So let's say he wants to increase the number of books available in the inventory. So that will mean quantity is going to be increased incremented by 50. So that's the UPDATE statement. And he's going to run that update. And we can look at the buyer one session, let's say buyer one is trying to buy the same book. And and then, so he's going to go on the website and then click on buy now or whatever and then is going
01:30:00
to translate into this UPDATE statement in the database, choose quantity equals quantity minus one. So reducing the quantity by one, meaning he's buying a, buying a book. And of course, there's going to be, you know, other statements updating other tables. But then to keep it simple, I'm just showing you the product table changes section. So as you can see, this is going to wait because seller is updating this particular row action. And that can be seen using acquittee. On data locks, so if you're under this greddy, of course, you can modify this query as per your needs. But then if you query this, you will
01:31:02
see that there's bunch of sessions and is, is the lock mode column. And then the table on which the database on which the locks are happening the table, so it gives you a lot of details. So, so if you want to understand what's going on here. So we have products table, and then we have ix lock, which is intention, exclusive lock on the table itself, meaning like a transaction is about to get an exclusive lock. And this is at the table level, but don't get tricked by that. There is also another row indicating there is a record level or a row level lock. And, and that is logging only this data equals one. So if you remember that UPDATE statement, we are using product ID. So and data for which
01:32:09
is one, actually, so product ID equals one. So that's what we are seeing over here. And if you see here, this buyer session has actually timed out already, so he's going to attempt to buy again. So that's how like, you can actually look at the locking details in this table. Let's try, let's say like buyer two comes in at this point. And then he just tries to browse the inventory on this ecommerce website. So that would mean a select query or read query. And he's, he's able to do this happy reaction. Right. So there is no problem. So while the rollouts are happening, other sessions can read this table, they can even look at the data for the same product. But they they just cannot buy this book, because that is being blocked by the seller. So again, it timed out. So at this point, buyer two wants to buy a different book, you know, I'm not able to buy this book, let me try buying
01:33:15
a different book, that's going to translate to, you know, product ID ID equals two, which is not being locked by the seller. And then that update goes through. And at this point, let's say the seller has completed updating the inventory. And, of course, if you look at the data, now, it's going to look different, because this has been updated to 150. And of course, this hasn't gone down because buyer, buyer, one is still in the process of buying the book, because the commit has not happened yet in the application. And then if we look at the data, again, the data has gone down, or the quantity has gone down, then via two, let's say wants to buy the first book that buyer one wanted to buy. At this point, there
01:34:16
are no no locks in this table. Because everyone's committed, and let's say buyer, who is trying to buy this, this book, and then he goes through with that date, and then commits and look at data. And then the data is changing actually. So this is how row level log basically allows for high concurrency. So only the rows which are logged by your transactions are not available for these other sessions to modify. Right So the other records which are not touched by your transactions are available for updating, deleting, etc. and all, of course, you can add new books, that means inserting new records in this table. So I just wanted to show you
01:35:22
the difference between table level logs and row level locks. So this session and my previous session will, will be useful in understanding that difference. Thank you, I'll see you in my next session. In this session, we're going to be talking about deadlocks. And I just want to show you how deadlocks happen, they do happen in in a busy ecommerce or B. database often, so it's good to know what they are. So it's going to be a very short and sweet session. So here, we have a couple of sessions again, so connecting a connected to the same database has two sessions, two different sessions. So let's say that we are working with products table, right. So we have seen the stable before in my previous sessions. Basically, this table
01:36:29
has information about the products that are being sold on, you know, an e commerce website. So we have a couple of records over here, you know, we're going to first let's say, you know, I seller comes to actually update the quantity of this product, basically, let's say if he wants to increase the quantity by 25. For this first book, this is the command that he's, you know, that's going to be executed, you know, whatever buttons he is clicking, will be translated to an update command like this. Right. And let's say like a different person from the same company wants to update the price of this book, not this book, let's say we have it the other book, I'm just actually using the product ID to update the right product, right. So we have one session where seller, one is updating the quantity of this item,
01:37:29
we have another session where we are updating the price of this item. And then if you see the prices incremented by two, let's say $2. And this is fine, right? So now we have row level locks. So this guy is holding a row level lock on this row. And this guy is holding a row level lock on this row. So this is fine, right? So we are operating on two different records, two different locks are independent of each other. All good. So now let's say the same seller, the second person who is updating with price, wants to update the price of this other book to actually like he is actually increasing the price. Again, by $2 of this book, the product ID equals one, which book, this one right here, let's go ahead and try to increment the price. By running this command, you know, he's waiting on waiting
01:38:38
for the lock ECI exclusive lock. And that's not available, because this seller has not committed actually is not committed. So let's actually go back here and, and this seller at the same time price to update the price of or quantity of this book. So two sessions are fighting for pretty much the same resource, you know, we ended up in a deadlock situation. So my sequel was smart, smart enough to just kill the session. Otherwise, we would have two sessions waiting for each other endlessly. Right? So here you can see the error code that is thrown, it says deadlock found when trying to get locks and try restarting that transaction. So let's go ahead and query the products table and see how it looks. You can see this, this whole transaction was rolled back. Correct. Both the transactions were
01:39:43
rolled back. There's even this one was rolled back. So I think that Locke was also killed. So that's why this this one went through. If you can see the prices have increased by $2. right because initially For 1699 and 2039, and here 8099 and 20 to 39. Okay, so that's how it works. This is a typical deadlock situation, I hope this explanation was clear. And I will see you guys in my next session. All right, my sequel learners. So in this session, we're going to talk about clustered indexes. So, so clustered index is not a different index type as in, like, you can, you know, directly create a create a clustered index yourself. So it is a type of index that, that MySQL kind of maintains in the, you know, behind
01:40:48
the scenes actually. So, in also your table data, the data that you insert into your tables or load into your tables are maintained in these indexes. indexes only what I mean by that is, so let's say this is a B tree index, right, so this is a B tree index. So you have my sequel, creating this B tree index, as you load the data into these tables. And then, you know, in the leaf nodes, what you have is actually the data, the data that you're loading into these tables, right? In the clustering, the sorting is based on the primary key that you define, or, you know, in this table, actually, so if you don't define a primary key, MySQL will automatically pick up a non nullable index key, what that means is, so let's say
01:41:49
that, in fact, actually, let's jump straight into the example that I have prepared for you guys. So so this is my MySQL Workbench. And, you know, I'll show you this table definition. So this is called products underscore one. And it's basically a products table that is typically used in a ecommerce store. And if you've been following my lessons, this is what I've been using, I just changed the name of the table for, you know, demonstrating this concept, this clustering, clustered index concept. So you have all these like columns, and I'm defining a primary key. Okay, so let's just start by, you know, I'm just going to switch to a database called eecom, store our schema called the econ store, I'm going to drop, you know, these tables if they exist already, by any chance. So the table doesn't
01:42:51
exist, which is okay, so I'm going to create this table, which I just talked about, called products. And then this table has primary key in a primary key is product ID. So product ID is sort of like an integer column. So this is an auto increment, right? So you don't even have to provide value for this column, actually, when you load the data, so you can just put all this information and load it and then we are good, MySQL will automatically increment the value of this column action. So and then, of course, like I said, like there is isbm column, which is over here, sort of leg book iasb. And information if you are, you know, if you remember your school days like this, this is be a number attached with any book, so something like that. So some kind of ISDN alphanumeric number. So
01:43:52
I'm going to call that like a unique key or a unique constraint. And let's go ahead and create the stable and this constraint. So that was successfully created. And I'm going to create a procedure, which I can use to kind of like populate the stable, right, so don't worry about the details of this procedure. This is something that I wrote to populate this table. And then that is successfully created and change the delimiter back to a semi colon. And then I'm going to call this procedure and which is going to throw some warnings, which is okay with me. As long as as long as the data gets populated, I'm fine. So it's going to probably generate some, you know, load some 6000 plus rows into this table. So we'll see how much we get this awesome. So it's actually loading a lot of data. It
01:44:54
seems to be done. So let's go ahead and commit the data and Now Actually, I'm going to select the data in this table, right? Just select all the data, and you will see that the data by default, or the data is actually sorted based on the primary key, which is product ID. And you can see, we know, I haven't like specified any ordering. So this is, you know, this is the default ordering of data, right. And so basically, your table data is sorted based on your clustered index, which is primary key over here, because you have the primary key in the table section. Right. So now the next thing is actually, I'm going to create a similar table, which is, you know, so I'm going to call it products too. But in this case, I'm going to basically not define a primary key, I'm still going to have a unique
01:45:55
key called, again, the same thing, you know, it's isbm, it's a unique key. And let's just give it a different name, just so we have kind of like, we have different names for different constraints. So let's actually go ahead and create this table. And so this table is created, I'm going to copy the data from the first table that you know, where I loaded a lot of data. So I'm going to copy the data from that table into this table, right. So just very simple. And then I'm going to commit, right, so that's a board 6455 6455 number of rows inserted into this table. And I'm going to select all the rows from this table. And you can see that now, the data is not sorted by product Id rather it is sorted by
01:46:56
this iasb. And it is sorting based on first character first, and then Initially, the first and second characters are the same, then 010 true. And that keeps going 05 and then 090, a BCD of GE hedge and then having after the zeros, you know, see one, so it is basically sorting data based on iasb. And and why is being because because of the absence of primary key, it's going to choose this iasb and column, as are the it's going to choose this non nullable unique index key, which is based on iasb and column, right. So it's starting based on this, but this is actually a terrible, terrible idea. Because if you're generating random, alphanumeric strings for iasb. And, you know, then you're not going to be generating the
01:47:57
string in sort of like an ascending order or in any type of order, actually. So in that case, actually, you know, when you're, as you're inserting data into the stable, this B tree is going to be created behind the scenes. And then my sequel, like whatever program is creating or maintaining this data structure behind the scenes has to work really, really hard to manage this Bre B tree index, actually, right. That's why this is a terrible idea to have like a you UID or some kind of alphanumeric string as a primary key actually, or in the absence of primary key. Well, my sequel is going to use this this key for clustering. And again, it is very bad. So keep that in mind when you're creating tables actually. Right. So finally, what I'm going to do is create another table called product three. And before that, I'm going to show you the output of this query, which is basically going
01:49:01
to come up empty or no, no, no road rows returned. All I'm doing is actually checking whether this index the index with named Jen flushed index is there in this database, actually. And then I'm checking the InnoDB tables and information schema I'm joining in odb tables and in odb indexes. And I'm checking whether this index indeed exists, right? Saying it doesn't exist, which is where the this, this credit, return no rows, and I'm going to create this table and this time, I'm not even going to create the create a unique key. And I'm going to make all these columns as nullable columns, you know. So I just want to show you what happens when you have a scenario where you're creating a table with all nullable columns and no primary key index no unique, not nullable index and you know Then I'm going to insert data into this table. Again, six, the 400 plus rows inserted, commit. And then
01:50:11
I'm going to select from this product three table right now. And when the data comes up, you can see that there is still some ordering that's happening. And, you know, we don't have any of these options primary key or a not nullable, unique key available, then how is MySQL able to sort data? What is it using, so it actually uses a hidden, hidden key actually, right, a hidden primary key. So if you run the same query, again, is ready, you can see that this index has been created on products three table, which is maintained internally by my SQL, for just the purpose of clustering this table, actually. Okay, so that's a lot of information. I hope you found this useful. And I will see you guys in my next video. Hello, my SQL learners. So in this session, I want to teach you the basics of using explain
01:51:18
or explained plan in MySQL. Alright, so now let's just let me just show you the table that I'm going to be working with, I'm going to be working with that table called products underscore one. And it's got some net in a product name, product type price. And if the product is a book, it will have an ISP a number attached with it. And then there is a quantity column. So these are some basic columns that you would see in an e commerce online store. So let's get started by just looking at the indexes of this table. So this basically has two indexes. One is a primary key index, which is on the product ID. And the other one is an index on the iasb and column. And this is a unique index, actually. So let's get started by picking a simple query that we are going to kind of like optimize using explain.
01:52:24
So the query that I'm going to be using is this. So I'm going to be selecting iasb. And from this products underscore one table where product name has cat in it. So the product name is cat. Okay, so And before I run this query, I'm going to look at the explained plan of it. And I'm going to put a slash g at the end. So I get that we'll put in there in a readable format. So first of all, it gives this output, right, and selectors, just one straights, simple select. That's what this is showing. But the main thing is we are working with our this particular row is referring to this table. And apart from that,
01:53:27
actually, you have all these columns, and then they are all null right now, like they don't make much sense apart from this. So this is a tight column and all means that it is doing a full table scan. Basically, MySQL is doing a full table scan, it's scanning the whole table. And how many rows is that it's these many rows. And we are using a filter over here, it gets all those rows and then it filters the output. And basically a you know, there's about 600 rows with product name equals cat, right, so the filtered person ages like 10%, basically, and then there is some extra information. Let's go ahead and create an index on this table. Create index called, you know, we can give an arbitrary
01:54:27
name. And, and I'm going to create add on products, one table and the column is product name, of course. This is the column on which I'm creating the index. Actually, let's just go ahead and run the explain again. So this is the explained plan. And that's how it looks. So basically, you can see that the Again, it's pretty much the same kind of output, but this time, it is also showing some data for all these columns. So first of all possible keys column shows like all the indexes that this query can use. And, and out of which, like this is the key or index that it is, you know, it is going to use this particular execution is going to use, and this is the key length in bytes actually write the number
01:55:33
of rows that is being scanned in this key, which is 589. And, you know, since this is index based, we're not really filtering data, rather, we're just going to the index and getting the data. So there is no filtering over there. Let's actually create another index, which also includes iasb. And, and see, like, what happens, actually, we're gonna create the other index and give it a different name. So let's go ahead and run the explain plan again. So now, again, the possible keys are these two indexes, but it still chooses to go with this particular index, and the index, key length is the same, and then grows,
01:56:35
and etc, etc. So there's no filtering that happened, right? Because we're choosing an index. So you might be wondering, like, you know, why it's not using the covering index, right. So this is supposed to be the covering index and covering indexes are supposed to be better than normal, non clustered index or a secondary index. So you can actually like, use a format like JSON format to get more information. So how you can do that is by just specifying like format equals JSON, and use the use that. And so that's going to give you the output in JSON format. And you can see that the you know, it gives you a little bit more information as then like the query cost, you know, this is how much
01:57:38
it's going to cost for my sequel to execute this query. And this is a representation of the amount of work MySQL has to do to run this query actually. So the cost for this one is 7690, right. And then again, it says these are the possible keys. And used key is used key parts is product name, which was not given over here. And then there is a cost and for which is a split of where the cost is going. So you can read my SQL documentation on all these fields. You know, you might be wondering why the covering index is not being used. And we can actually force that index by using this use index. Syntax or use index keyword. And then I'm going to put the index name that I want to force which is this one.
01:58:46
And when I ran it, this ran the explained plan this time, it shows the cost of this one is going to be 109 point two seven, you know in comparison to the previous explain plan, where the cost is only 76. And this is why my sequel is going with this particular plan instead of this guy. Okay, I hope this session was useful.

DOWNLOAD SUBTITLES: