用 SQL 找出某只股票连续上涨的最长天数

涉及多张中间表:

SELECTMAX(consecutive_day)FROM(SELECTCOUNT(*)as consecutive_day
  FROM(SELECT trade_date,SUM(rise_mark)OVER(ORDERBY trade_date)AS days_no_gain
     FROM(SELECT trade_date,CASEWHEN closing_price > LAG(closing_price)OVER(ORDERBY trade_date)THEN0ELSE1ENDAS rise_mark
           FROM stack_price) subquery1) subquery2
  GROUPBY days_no_gain) subquery3;

Over 语法

SELECT
  product_id,
  sale_date,
  sale_amount,SUM(sale_amount)OVER(PARTITIONBY product_id ORDERBY sale_date)AS total_sales,SUM(sale_amount)OVER(PARTITIONBY product_id)AS running_total
FROM
  sales;

basic:

order_id | customer_id | order_amount
-------------------------------------1|1|1002|1|1503|2|2004|2|505|2|120

result:

order_id | customer_id | order_amount | total_amount | running_total
-------------------------------------------------------------------1|1|100|100|2502|1|150|250|2503|2|200|200|3704|2|50|250|3705|2|120|370|370

Window function

A window function is a type of function in SQL that performs calculations across a set of rows called a “window.” The window is defined by the OVER clause, which specifies the partitioning and ordering of the rows.

SUM(order_amount) OVER (PARTITION BY customer_id ORDER BY order_id):

SUM(order_amount): This is the window function itself, in this case, the SUM function is used to calculate the sum of the order_amount.
OVER: It introduces the window function and specifies the window’s characteristics.
PARTITION BY customer_id: This clause divides the rows into separate partitions based on the customer_id. Each partition will have its own calculation of the sum.
ORDER BY order_id: This clause determines the order in which the rows are processed within each partition. In this case, it orders the rows by the order_id.
SUM(order_amount) OVER (PARTITION BY customer_id):

This is another usage of the SUM window function, but without specifying the ordering using ORDER BY. Without the ORDER BY clause, the entire partition is considered, and the calculation is performed on all rows with the same customer_id.
The window function, in combination with the OVER clause, allows us to perform calculations within specific partitions and orderings defined by the columns specified. It provides a way to aggregate or calculate values based on a subset of rows without collapsing the result set or using subqueries.

Other common window functions include ROW_NUMBER(), AVG(), MIN(), MAX(), and LEAD()/LAG(), among others. Each function has its own specific purpose and behavior within the window frame defined by the OVER clause.

OLAP / OLTP

SQL 作为查询语言而发明, 名字叫 “结构化查询”（structured query）, 数学基础是 “关系模型”, 没有考虑复杂计算 (与之相对的是离散数学, 把 “数据存储 + 数据计算” 做在一起)

近年来, 数据处理和计算的需求越来越大, 于是 OLAP（联机分析处理）和 OLTP（联机事务处理）的概念就诞生了.

• OLAP: Online Analytical Processing.
• OLTP: Online Transaction Processing.

它们基于数据库, 属于"数据库 + 计算层".

处理海量数据, 有效率瓶颈.