You need to create the dataset.
Which dataset mode should you use?
You need to create the dataset.
Which dataset mode should you use?
To create the dataset, the Composite mode should be used. This mode allows for a combination of Import and DirectQuery, which can provide the flexibility needed to handle different data sources efficiently. Composite models are particularly useful when dealing with large datasets that may be too extensive to import in their entirety, but where certain portions of data need to be readily accessible for fast querying. This mode also supports configuring the storage mode for each model table, optimizing the balance between performance and storage requirements. Given the diverse data sources and the need to meet specific report and technical requirements, Composite mode offers the most comprehensive and adaptable solution.
You wouldn't use composite for all. I would say import as the SQL Server data is only 2GB and excel is really small. Also, only need it refreshing once a day so this dataset is very small. Answer is A (Import)
Capacity for import is 1 GB, so How can we use Import for 2GB data
I think the size of the data is 2GB in the database though not Vertipaq compressed. 1GB limitation applies to size of the pbix file which is already compressed and will be significantly less than raw data.
I worked with DBs that are much larger than 5GB and have always used import due to fast response/RLS/Time Intelligence etc...There are a few ways to limit the dataset size, most common is to use SQL with Where clause to limit the data to only that you needed or if you don't use/don't know SQL, do a transform before loading the data. I would do an import in this case but my answer will not be neccessarily the correct answer that Microsoft wanted.
it is, for excel files
what do you mean by "all"? The technical requirements state that all reports must be supported by a SINGLE dataset and that one will definitely exceed the 1GB limit.
Also data is 2gb and 5mb but we need only top 20 reports.
you missed this detail : The warehouse shipping department must be notified if the percentage of late orders within the current month exceeds 5%.. How do you put an alert then ?
Composite Model means now you can have a model, that very large tables of that are coming from the DirectQuery connection, without the need for importing, and small tables to be imported to be accessible quickly. In this case, Composite fits perfect. Source1 is 2GB, which is relatively too large for daily updates and way larger than Source2 which is only 5MB.
Report data must be current as of 7 AM Pacific Time each day The reports must provide fast response times when users interact with a visualization
There are 2 sources. first is azure sql data base with 2GB of data and other one is excel 5 MB. You can not load 2GB of data in model hence you have to use direct query and excel we can import. Hence it is composite mode
A: I believe that since they specifically call out a time of 7AM its either a trick lol or its intended to tell us IMPORT is fine (since its 7AM) Secondarily, they talk about performance and since you can import that easily, the performance will be much better on IMPORT and DirectQuery is known to induce latency / lower performance. So I don't see a reason to select Composite in this case.
A. Import The Import mode is the best choice here. This mode allows you to import the data into Power BI's memory, enabling fast response times when users interact with a visualization, which is one of the requirements. DirectQuery and Live Connection modes would not be suitable because they keep the data in the source system and query it from there, which can lead to slower response times. The Composite mode, which allows a combination of Import and DirectQuery, is not necessary here as there's no requirement to combine these two modes. Remember that with Import mode, you need to refresh the data to keep it up-to-date. The requirement is to have the report data current as of 7 AM Pacific Time each day, which can be achieved by scheduling a daily refresh in Power BI.
What about row level security, I believe it (RLS) does not work well with DirectQuery. (A)
I'm confused by the case studies. Are we going to be interacting with a user interface or are we just selecting a multiple choice answer or is it both?
You wont have a UI unless you mean possible drag and drop answers. Then it would be both. If you meant actually like the SMS / Admin / Etc / Gateway, Power BI, no you wont touch those.
Near the real-time data is not required. 2GB size data seems large but it is the whole raw data, we need the latest 3 years data only here and we can pre-aggregate it. So, I prefer "A" which gives the highest speed and flexibility here.
2 different sources = Composite Model
C is the correct answer. I will go for the composite model. Bear in mind that the capacity for import mode is 1GB only.
how do you manage : The warehouse shipping department must be notified if the percentage of late orders within the current month exceeds 5%. in a full data import ? As far as I'm concerned, For that issue, I would put an alert on data change so tell me how you will put an alert in your "light for now" early morning import ? :-)
I had a hard time choosing between import and composite. Ultimately, what another user posted reminds us: the database is 2GB+ of text and growing. Do we want to be running a daily import on that? I think this is why the given answer is composite.
Source1 has a datasize of 2gb. The text does not specifically state that there is a premium capacity, so I suspect that the model size for an import model is 1GB. Is there a way to know how big this model size is?
Composite mode because it can mix Import and DirectQuery modes.
I think is Import. Because Direct query cannot fulfill this "Report data must be current as of 7 AM Pacific Time each day.", it will pull the latest lets say if it is 10AM, it will pull 10AM data right? Please correct me if I am wrong.
I also go for A. Import mode meets the requirements for fast response times and current data at 7 in the morning. Though the model must minimize the size of the dataset "as much as possible", it must meet the report and technical requirements. Fast response time is a hard requirement and it's sure that Import mode delivers that better than DirectQuery. I am no expert on Composite but I would argue that in the best case, it's as fast as Import mode. In Composite mode, some tables might switch to DirectQuery due to the user behaviour, but that would mean we would overachieve the requirement for daily refreshing, buying this overachievement through slower response time. That's not what we want. Future growth of the data is not to be considered in a question like this.