Why is this inner join necessary when using same table? - mysql

Question

Why is an inner join necessary necessary when all info is located in ONE table? Explain please. just getting hang of joins.
If you were interviewing for an entry level analyst job (which I am), given the following sample query and table description, what other observations/questions would you have on this information?
Table: sales_data
**Field** **Description**
date Date of order|
customer_id Unique ID for Customers|
brand Unique ID for the brand of a purchased item|
quantity Quantity of item purchased|
amount Item Price
select customer_id, brand, sum(quantity*amount) as sales
from sales_data a
inner join (
select distinct customer_id
from sales_data
where lower(brand) = 'chanel'
and quantity > 0
and date >= to_date('01/01/2017','MM/DD/YYYY')
) b
on a.customer_id = b.customer_id
where date >= to_date('01/01/2017','MM/DD/YYYY')
and quantity > 0
group by customer_id, brand

Answer 1

Inner join in the same table sometimes required if you fill all your data in one single table with redundant information. For the table description (You explained), there could have Three different table easily - customer, brand & sales_details with appropriate relational key. And that case no Self Join (inner join in the same table) not required. But as data is available in a single table, the requirement actually lead you to the necessity of using a Self Join query. As per output of the query, its clear that you need this following requirement steps-
First you need a distinct list of customer who are selling product with brand='chanel'
Secondly, you need all product selling details for those selected customers in step one.
To avail results for above requirement, a Self Join is required.
Other Question: How can you re-write the query avoiding Inner join to pic the same result?
Ans: You can simply use a inner query in the where statement as below to get the same result-
select customer_id, brand, sum(quantity*amount) as sales
from sales_data a
where date >= to_date('01/01/2017','MM/DD/YYYY')
and quantity > 0
and a.customer_id IN (
select distinct customer_id
from sales_data
where lower(brand) = 'chanel'
and quantity > 0
and date >= to_date('01/01/2017','MM/DD/YYYY')
)
group by customer_id, brand

Answer 2

Your sales_data table presumably contains data for all of the sales of customers. So lets say you wanted to ask a question such as:
What is the total sales revenue for customers that have purchased 'chanel' brand?
The first thing you need to do is identify all of the customers that have purchased 'chanel' then you need to go and get all of their other sales to determine the total revenue.
The inner/sub query does exactly that it identifies the customers that have purchased 'chanel' brand since 1/1/2017.
select distinct customer_id
from sales_data
where lower(brand) = 'chanel'
and quantity > 0
and date >= to_date('01/01/2017','MM/DD/YYYY')
The outer query then joins to those customers and gets all of the sales data by brand by customer since 1/1/2017 that have a positive quantity.
select customer_id, brand, sum(quantity*amount) as sales
from sales_data a
inner join (
Customers That Purchased 'Chanel'
) b
on a.customer_id = b.customer_id
where date >= to_date('01/01/2017','MM/DD/YYYY')
and quantity > 0
group by customer_id, brand
There are methods of using EXISTS or IN to accomplish the same thing. However be very careful of IN if a customer_id could be NULL.

Answer 3

Yes, you are right you can do with one referencing to the table. The inner join gets particular rows and then calculated their sales value. You can do this directly:
select customer_id, brand, sum(CASE WHEN and lower(brand) = 'chanel' THEN quantity*amount ELSE 0 END) as sales
from sales_data
where date >= to_date('01/01/2017','MM/DD/YYYY')
and quantity > 0
group by customer_id, brand
having MAX(CASE WHEN lower(brand) = 'chanel' THEN 1 ELSE 0 END) = 1;