Exam PL-300 topic 2 question 35 discussion

My answer is NO NO YES Summarizing Orders by CustomerID, OrderId and OrderDate means to group by CustomerID, OrderId and OrderDate and to aggregate the rest of the fields, however the OrderId has unique values so the aggregation will have the same number of rows as the original table

Box1: Yes The summarizing data will remove Product information which is not required for this analysis. We should not infer from the sample data that OrderId is the primary key for few reasons. 1) the Order table is a common table that is used by Microsoft in learning materials, and the table usually has a surrogate key as the primary key. 2) Order ID can be printed on the invoice, and it is not a surrogate key. BTW, the Customer ID is not a surrogate key as well. 3) An order usually consists of multiple products. 4) The Product Id is unique in the sample data as well, and we won't infer that we must create an order for a separate product. Box 2: No Customer ID is required for new and returned customer analysis Box3: Yes Product information is not required for analysis. Note: if the first analysis requirement changes from "Total sales over time" to "Total sales over time for products", we will have a different question to deal with

NO NO YES

I would vote for NO - NO - YES. I was thinking YES for the first one as I cannot see how this summarization could affect required nalysis but then, after reading some comments, I realized this summarization would not reduce the data size as all three fields using summarization looks like unique values to me, therefore, no row reduction would result from this summarization

Since OrderId is unique for each order, summarizing the Orders table by CustomerID, OrderId, and OrderDate will not lead to a reduction in the number of rows or the model size. Therefore, it won’t have a meaningful impact on the dataset.

In my opinion it is YES, NO, YES

Yes: Since you don't know the order ID is unique or not but it also contains products so probably not otherwise would have been orderline ID. YES: You can count the new customers based on this table so no need for a relationship with the customer, there is no extra information visible within the customer table. YES: Don't need the columns

Yes No Yes

The correct answer is : * No ( "Orderid" is unique so can't be aggregated with other two columns) * No ( Star schema, "Customerid" is needed) * Yes ( Don't need those columns for analysis)

When you look at the data in the table you see customer TORTU having ONE order composed of THREE different productID with the same ordered date. Furthermore, looking at the data of OrderID, it pretty looks like an autonumber to me which is unique and incremental. Some might say no, TORTU made 3 different orders. Okay but then I suppose ordered date should be different, right ? Based on those assumptions, summarizing will have no effect on the size. So, NO, NO, YES

No / No / Yes

1/ yes because we remove ProductID. impossible to know if it is unique like OrderID 2/ No, CustomerID is used in a relationship 3/ No, UnitPrice is useful for calculating total sales. I don't understand why so many people think it's Yes. How do you calculate Total Sales?

1/ oui car on enlève ProductID. impossible savoir s'il est unique comme OrderID 2/ Non, CustomerID est utilisée dans une relation 3/ Non, UnitPrice est utile pour calculer le total des ventes. je ne comprend pas pourquoi il y a autant de gens qui pensent que c'est Yes. Comment calculez vous Total des ventes ?

NO NO YES No: Summarizing orders with these columns may actually increase the model size as creates a more detailed representation of the data. It will not reduce the model size. No: Since there's a relationship between the Orders table and the Customers table using the CustomerID column, removing it might affect the ability to analyze data by customer, so it may not support the current analysis. Yes: Removing unnecessary columns like UnitPrice and Discount that are not used in the analysis will likely reduce the model size without affecting the ability to analyze total sales over time, order counts, and customer counts.

Fer079 8 months, 4 weeks ago if you see the table, the order ID is unique, so you are not going to have the same Order ID more than twice. For example, you will see that for the customer TORTU has 3 different products (18, 63, 75) under the same order because the OrderDate is exactly the same so we have to think it´s the same order, however the OrderID is sequential (unique). So NO, NO, YES

My answer: NO: orderID is unique, suppose s CostumerID has several Orders in a day. we cannot aggregate that. NO we need CustomerID for the relationship YES we do not need them.

1st statement NO. TotalSales is not included in the summarization, therefore won’t support current analysis. Normally after creating the aggregated table, you would disable the original table from loading into the model to reduce overall size. 2nd statement is NO, you can’t remove the key column, on which relationship is based. 3r statement is YES, you can remove those columns as they are not needed for current analysis. TotalSales will serve the purpose to calculate Total sales over time. I assume TotalSales is a normal column, not calculated one.