Harrow Data Analysis (SQL) Written Exam

data

import database

First import the database:

import separately:






as follows:

import several other data tables in the same way:

understand data

Category: cate Brand: brand Sales: qtty Sales: amount User ID: pin Product ID: SKUID

• Pin table : three characteristic relationships of dt, skuid, and pin. Namely: time, user ID and item ID.
• sku table : the relationship between skuid, cate, and brand features. That is, the relationship between the product number, category, and brand.
• Order table : The relationship between dt, skuid, sku_name, qtty, and amount features. Namely: time, commodity number, commodity name, sales, sales.

A category is a category under a brand. Such as Huawei laptops.

Require

1. All products with a unit price of 0 are not included in sales . 2. Notebooks with
a unit price of less than 500 are not included in sales
.
The number of de-duplicated users of the brand under the category (for example, a user id browses three products of Xiaomi in a single day, the number of visitors to the Xiaomi brand under this category is recorded 1)
5, 2019 and 2018 sales are the sales of No.
1-156 、 Year- on-year data can be presented in
decimal form, keep two digits

ideas

• The order table is associated with the sku table to take out the brand of the order
• Use if in sum to judge whether it is 19 years old
• The first two are where filtering, followed by case when renaming, first filtering in the subquery, this mainly examines the code style

my code

SELECT
	result2.*,
	ROUND( result2.19年销量 / result2.18年销量, 2 ) AS 19年销量同比,
	ROUND( result2.19年销额 / result2.18年销额, 2 ) AS 19年销额同比,
	ROUND( result2.19年访客量 / result2.18年访客量, 2 ) AS 19年访客量同比 
FROM
	(
	SELECT
		result.cate 品类,
		result.brand 品牌,
		SUM(
		CASE
				
				WHEN YEAR ( result.order_dt ) = 2019 
				AND result.unit_price > 0 
				AND result.sku_name NOT LIKE '%补差价%' 
				AND DAY ( result.order_dt ) BETWEEN 1 
				AND 15 
				AND ( CASE WHEN result.cate = '笔记本' AND result.unit_price < 500 THEN NULL ELSE result.qtty END ) THEN
					result.qtty ELSE NULL 
				END 
				) AS 19年销量,
				SUM( CASE WHEN YEAR ( result.order_dt ) = 2019 AND result.sku_name NOT LIKE '%补差价%' THEN result.amount ELSE NULL END ) 19年销额,
				COUNT( DISTINCT CASE WHEN YEAR ( result.pin_dt ) = 2019 AND result.sku_name NOT LIKE '%补差价%' THEN result.pin ELSE NULL END ) AS 19年访客量,
				SUM(
				CASE
						
						WHEN YEAR ( result.order_dt ) = 2018 
						AND result.unit_price > 0 
						AND result.sku_name NOT LIKE '%补差价%' 
						AND DAY ( result.order_dt ) BETWEEN 1 
						AND 15 
						AND ( CASE WHEN result.cate = '笔记本' AND result.unit_price < 500 THEN NULL ELSE result.qtty END ) THEN
							result.qtty ELSE NULL 
						END 
						) AS 18年销量,
						SUM( CASE WHEN YEAR ( result.order_dt ) = 2018 AND result.sku_name NOT LIKE '%补差价%' THEN result.amount ELSE NULL END ) 18年销额,
						COUNT( DISTINCT CASE WHEN YEAR ( result.pin_dt ) = 2018 AND result.sku_name NOT LIKE '%补差价%' THEN result.pin ELSE NULL END ) AS 18年访客量 
					FROM
						(
						SELECT
							sku.cate,
						CASE
								
								WHEN `order`.sku_name LIKE '%荣耀%' THEN
								'华为' ELSE sku.brand 
							END brand,
		`order`.skuid,
		`order`.qtty,
		`order`.amount,
		`order`.sku_name,
		`order`.amount / `order`.qtty AS unit_price,
		`order`.dt AS order_dt,
		pin.pin,
		pin.dt AS pin_dt 
	FROM
		sku
		JOIN `order` ON sku.skuid = `order`.skuid
		JOIN pin ON sku.skuid = pin.skuid 
	) AS result 
GROUP BY
	result.cate,
	result.brand 
	) AS result2;

result: