Hello and welcome back to another exciting episode on the tech blackboard. In the last episode, we looked at the details on DP900. What is DP900?
Who should do it and how you can get a free Microsoft exam voucher for the same. We also discussed upon what kind of questions should you expect from DP900 exam. The link for that episode is now appearing in the I button above on the top right corner and also in the description box below.
In this episode, we are starting with DB900 sample questions. It's important to understand that I have deliberately repeated some of the questions and presented them in slightly different format to help you counter any variations that might appear in the exam. Do not miss my another exciting series on A900 or Azure fundamentals.
The link for the same is also shared in the description box below. [Music] So let's start with our first question. Our first question is descriptive analytics tells you and then there is a drop-own menu given here.
And from this drop-down menu, you have to choose one of the options so that this statement is true. So now let's see the four options given in the drop-down menu. The first option is what is most likely to occur in the future.
The second option is what occurred in the past. The third option is which actions you can perform to affect outcomes. The fourth one is why something occurred in the past.
Now let's understand what kind of different analytics do we have before we answer this question. So the first one is descriptive which is given here as well. The descriptive analytics tells you what is happening or the descriptive analytics also tells you about what has happened in the past.
Then the second analytics which is diagnostic analytics tells you why something is happening. So basically you are trying to diagnose something right you're trying to examine something and then that's why it says why something is happening. The third is predictive analytics.
So predictive is what will happen in the future. So predictive is more related to examining the future. Thinking about what will be the outcomes in the future.
The fourth type of analytics is prescriptive. So prescriptive tells you what action should we take. So basically prescriptive is like something like uh uh let's say you go to a doctor and the doctor gives you a prescription, right?
So what is a prescription? Prescription tells you what you need to do, which medicine uh you need to take, right? So that's a that's more of a prescription.
So this is how you can relate prescription with prescriptive. So prescriptive is basically what action uh should you take. So now you understand there are four kind of analytics descriptive, diagnostic, predictive and prescriptive.
And as I told you that descriptive analytics tells you more about what is happening or what occurred in the past. Thus the correct answer for us is yes. What occurred in the past?
Okay. So now with this understanding let's move to our second question. Our second question is related to the batch processing.
Before answering actually this question let's understand what kind of uh processings do we have. So first one is batch processing and the second one is realtime processing. The core difference between the two is when you talk about batch processing you're talking about when you process the data in chunks or in batches.
Right. Um, and when you talk about real time processing, you process the data as and when it comes. So the records are coming and you're processing them as and when they are coming.
Right? So that's the big difference between batch processing and parallel or realtime processing. Sorry, not the parallel but the real time processing.
Okay? Now let's with that core understanding between uh the two batch and real time. Now let's try to answer the question.
So what are the options given? The options given are data is always inserted one row at a time. No, this is not related to batch processing.
The data is processed in real time. Definitely not. You can already see the keyword here which says real time.
Definitely it's related to the realtime processing. Latency is expected. Now this looks very close to the answer.
Why? Because as I told you batch processing is when you process the data in big chunks right so when you're of course processing the data in bigger chunks of course the processing will need some time and that's why latency would be expected. The fourth option is processing can only execute serially.
No uh this is not the right answer because in batch processing you can have two modes serially or parallelly. So you can choose any of those. So you don't need to depend actually on C relay.
So the correct option for us is latency is expected. Okay. With this now let's move to our third question.
Our third question is related to the extract transform and load process which is in short form is also called EDL process. I think you might have already heard if you have any time done some work with uh extraction building some pipelines um then you must have heard this term ETL process so now let's look at what is ETL process so in ETL process you basically extract data from some source transform the data with some business logic on this and then load the data into some data store right so Now um let's look at the options here. A matching schema in the data source and the target.
A target data store powerful enough to transform the data data that is fully processed before being loaded to the target data store. Or the fourth option given is that the data target to be relational database. Now if you already know about ETL process you would also know that in the ETL process what we are doing is more importantly we are transforming the data before it's getting loaded into a data store.
So when you're transforming the data which means what you are already processing the data you're filtering the data you're cleansing the data before it's actually loaded into the data store. So with that understanding I hope you can already answer the data uh question. So the correct answer for this question is data is fully processed before being loaded with the target data store because we are already transforming the data.
So this is very important always always check out this middle keyword is it transform or is it the load. So in the next question it's a slight variation of this same process. So this is ETL process and in the next question four we will look at the ELT process.
So this is our question number four. Now you see there is a variation of the same question and Microsoft or in the exam you might be asked these kind of very similar questions to confuse you but do not get confused just understand the concept and you will never make a mistake. In this question, you are asked about the extract, load and transform.
I already told you before that always look at the middle part. Is it a load or is it a transform? So in this process uh the ELT process, it's a load which is getting before transforming.
So when you load the data into something, what would you need? Now you're loading a big chunk of data, you would need something powerful to process the data, right? So that's why if you will, if you think with that understanding and read the options now, a separate transformation engine, no, we don't need that.
A target data store powerful enough to transform the data or the data is fully processed before being loaded to the target schema or the target store. Or the fourth option is a data pipeline that includes transformation engine. I already told you when you're loading the data before transformation then you need something powerful to process the data, right?
So that's why our correct answer is a target data store powerful enough to transform data. Now let's move ahead to our fifth question. Our fifth question is related to the relational databases.
So it says a relational database is appropriate for scenarios that involve a high volume of and your options are changes to the relationship relations between entities, geographically distributed rights. The third option is transactional rights and the fourth option is rights that have varying data structure. If you have already read about relational database, you probably can already guess the answer.
But I will take you through. So the first option is changes to the relational relationship between the entities. Guys, when we are talking about relation uh relational databases, we know that we don't want to change the relationships between the entities.
They are already predefined. We have already, you know, checked the business logic and based on that we have built the relationships between the entities. and then we don't want to change them.
So definitely this is not an answer. Then geographically distributed rights though definitely not it's very difficult to actually geographically distribute the relational database. It's it's actually very fit for the non relational database.
So definitely this is not the answer. Transactional rights. Now this looks interesting.
Why? Because most of the relation data databases are actually fit for OLTP right? You know this word OLTP OLTP which means what?
Online transactional processing. Now you can already relate it with this with this third option OLTP. So we are already talking about transaction.
So OLTP are suited for relational database. So thus this looks to be the answer. But let's look at the fourth option as well.
The fourth option says rights that have varying data structure. Definitely not. This is a big no.
Why? Because when we are talking about relation database, we always set the data structure well in advance. They are well thoughtful.
Uh they're well planned and the relationships between them are already well set. We don't change them during the course of processing. So definitely this is not the answer either.
So our correct answer is transactional rights. Okay, I hope you're liking the questions. Okay, now let's move our next question which is question number six.
The question number six says about transcribing audio files is an example of dash analytics. I already talked about descriptive, predictive, prescriptive and the fourth type of analytics which was diagnostic. Right?
So we already talked about these kind of diagnostic in the first question. Right? So now you understand that descriptive is what's happening, diagnostic is why happening, predictive is what will happen and prescriptive is what action should we take.
Now transcribing audio files, what does that actually mean? You need to understand this first. Transcribing audio files means that you are generating text, right?
You input a audio file and you generate text for it. That's actually what it means by transcribing audio files. So then it has nothing to do with these three definitely not.
So the obvious choice which is left with us is cognitive analytics. So let's move to our question number seven. Our question number seven is wait a minute you are already thinking that this question is repeated right?
Yes, this is but I already told you in my opening slide throughout this series I will be repeating questions not only repeating I will be presenting them with slight variations right so because in the test in the actual Microsoft test you will be getting similar questions which might have same answer however the way they are presented in the exam will be little different so you don't confuse that's why I'm giving the same questions in a repeated format or in some variations. Right? So let's continue and move ahead with So let's continue and move ahead with the question number seven.
So relational database is appropriate for scenarios that involve high volume of now you already I already told you that relational database are well suited for OLTP, right? So, OLTP is online transaction processing. That's why the correct answer is transactional rights.
Now, let's move quickly to our eighth question. Our eighth question is a visualization that shows a university's current student enrollment versus the maximum capacity is an example of is it cognitive, descriptive, predictive or prescriptive. So now I already told you in the first question or the sixth question in the last slide that there are four forms of analytics.
First one is descriptive, predictive, prescriptive and diagnostics. Right? Now if you read the question very carefully, you know that this is what we are talking about current student enrollment.
And I already told you that descriptive analytics tells you two things. Either it tells you what's happening now or it also tells you what has happened in the past. So when we are talking about current student enrollment, we are talking about something which is happening currently now.
Thus the correct answer for this question is descriptive analytics. Let's see the question number nine. Our question number nine is again related to relational database.
So relational database must be used when the options are a dynamic schema is required. A data will be stored as a key value pair storing large images and videos. Strong consistencies guaranteed guarantees are required.
Now if you talk about relational database there is no dynamic schema. I already told you in the last slide the schema the table structures the relationship they are already preset pre-planned predetermined you are not going to change them frequently in relational database it's not suited for that uh sort of transactions right so definitely this is not the right answer data will be stored in key value pairs definitely not this is something else we we normally use other type of databases like Azure tables for these kind of purpose purposes. Then we have storing large images and videos.
Definitely relational database is not suited for suited for this because if you want to store uh large images or videos, you use other options like Azure blob storage. Right? So this is not suited.
Then we have stoing strong consistency guarantees are required. Definitely this is the right answer because when we're talking about relational database the main crux or the main purpose of relational database is that we have a strong consistency of the transaction that are being done. So either the transaction is fully committed or the transaction is rolled back.
That's that's actually the meaning of consistency of the transaction. That's why this is the correct answer. Now let's move to the question number 10.
Our question number 10 talks about dash natively supports the analysis of relationship between the entities. We already talked about this isn't it? I hope you remember this that what natively oh okay so now let's look at the option the options are column family database document database graph databases or key value databases so what do you think the answer should be so guys when we're talking about relationship between the entities ities column family is not the right answer definitely not then we have document database no they are more suited to store the documents uh like JSON documents or let's see do we have graph database well graph database are actually very good for storing relationship or showing the hierarchy between the entities so this looks to be the right option but let's see the last one key value store they are absolutely not related to uh storing the relationship between the uh entities.
Thus the right option here is the graph databases. Now let's look at the question number 11. Question number 11 is related to relational data.
So relational data uses what? Uses what to enforce relationship between different table. So I already told you in our last slide that in case of relational databases you have tables in tables you have columns and on columns you you set primary keys or foreign keys.
So the primary keys and the foreign keys help you create relationship between the tables. Thus our correct answer is keys. So you use keys to enforce relationships between different tables.
Now let's quickly move to our question number 12. The question number 12 says creating closed caption text for audio files is an example of is it cognitive, descriptive, predictive or prescriptive. If you remember in the last slide also we had a similar question which said transcribing audio files.
So both actually means the same thing. They both mean that you are generating text for audio files. So do not get confused by different variations of the same question.
The answer remains the same. And yes, you guessed it right. It is cognitive analytics.
Let's move to our question number 13th. Our question number 13th is a little different. In this question, we are given a JSON file and there are questions based on this JSON file.
Let's look at the structure of JSON file before we look at the questions. So our JSON file has details of a customer. It has its first name, its last name, address, then it gives its social media.
Then it also gives information about phone number. Now let's look at the questions which are based on this JSON file. The JSON file here and the questions are here in this segment.
So the question says what is a customer? Is a customer a nested array, a nested object or a root object? If you have already dealt or worked with a JSON file, you would already know what is a customer.
But let me tell you and explain this JSON file and the structure of the JSON file a little bit more. So whenever you are dealing with the JSON file or you get a question based on a JSON file, always look what is the outermost node. In this JSON file, you can see the customer here.
There is the opening and this closing bracket, the curly braces or curly bracket as you want to say it. And then everything else, all the other information is actually inside these two curly braces. All right?
So what does that mean? That means that customer customer is a root object. Right?
And that's the answer for this first question. So customer is a root object because everything else is lying under it. So it's a root for everything.
Right? Now let's move to our second question. So what is an address and what is a social media?
Okay. So now before I tell you what is a address, let me tell you what is a social media right. So if you look at social media, social media is all is is starting with this square bracket and here it is ending with a square bracket.
So whenever you see a node in a JSON file which starts with a square bracket and ends with a square bracket, you always know that this is a this is a nested array. Right? So array in a JSON file is represented by the square brackets.
So what what is the address then? An address is definitely then is a nested object. Right?
So here you can see you can see the address here. Here sorry. So here we have a address.
So address doesn't have the square brackets. So we already know that it's not a nested array. It's also not the root array because it's inside one more node.
So it's not a root array as well. So the obvious choice we are left with for address is a nested object indefinitely. Then for the third part which is social media.
The answer is the nested array. This concludes our first part of DP900 questions. In the next part, we will take more such interesting questions.
Don't forget to like and subscribe and share your valuable comments. Until we meet again in our next video, stay fit, keep learning and thanks for watching.
