a select statement
basic grammar
select column name 1, column name 2 //You can use fully qualified column names tables. column names
form tables
filter(where)
Grouping (group by column name 1, column name 2) // The grouping basis must be all the column names of the select (except for statistical calculations), otherwise there may be ambiguity. See below for details
Having filters the grouping, not a mission alias (alias defined in select)
order by
limit clause
1) distinct keyword
select distinct A,B,C from tables; From the tables in the table, query the three columns of ABC, distinct indicates that at least one of the three columns of ABC is different.
2）LIMIT
select A,B,C from tables limit 5; display the first five rows
select A,B,C from tables limit 5,5; display the five rows starting from row 5, the count starts from row 0
3) order by clause, sort the index rows
order by A,B sort by A first, then sort by B The default is ascending order
order by A desc,B is sorted in descending order by A, and then in ascending order by B, desc only modifies the column in front of it.
3) where clause
= != < <= > >= BETWEEN specifies between two values
select A,B,C from tables
where c between 2 and 10; // for assignment columns can be indexed like this where c is NULL
select A,B,C from tables
where A<2 AND C>10 ; // OR also works AND has higher priority than OR A = 1 OR A=2 AND C<5 is equivalent to A = 1 OR (B>2 AND C<5)
IN is used to specify a range of conditions
select A,B,C from tables
where A IN(1002,1003) // IN clause can also contain other SELECT statements
order by A
4) LIKE text matching
A like 's%e' contains any arbitrary character, that is, only lines like s....e are matched, as.....e cannot be matched, LIKE is a full match
A like 's_e' underscore means any character
5) Regular expression REGEXP (regular expression rule, written later) Substring matching
A REGEXP '1000' All columns in column A that contain the substring '1000'
1000|2000
[123] 1 or 2 or 3
[0-9] [az] range or [:digit:] any digit same as [0-9]
The 4 after [:digit:]{4} indicates the match four times, that is, any 4 digits
[:digit:]{4,} no less than 4
[:digit:]{4,6} between 4 and 6
^start of text ^[0-9\\.] matches a number or . at the start of text
$ end of text
[[: start of word
[[:>:]] end of word
6) AS and concat function
concat(A, '(', B,')') output column name is A(B)
A AS E gives the column an alias E for display output
7) Arithmetic calculation (+ - */)
select procid, price, num, price*num AS total
form tables;
where total>2000;
order by total; output the id unit price quantity and total price information of all products with a total price greater than 2000 and sort them
8) Function
text manipulation functions
left() returns the character to the left of the string
length() the length of the string
lower() converts to lowercase
LTrim() removes spaces from the left of the string
RTrim()
Upper()
select procid,Upper(procname), price, num, price*num AS total //Upper(procname) output the product name in uppercase
form tables;
where total>2000;
order by total;
Date and time processing functions Numerical processing functions (absolute value, trigonometric functions, etc.)
9) Statistical functions
AVG AVG（distinct pro_price）
COUNT count
MAX returns the maximum value of the specified column (number or date), if used for a text column, returns the last row in positive order (sorting first makes sense)
MIN
SUM sum of column values
select AVG(pro_price) AS avg_price returns the average price for a specific column
from tables
where pro_id =2001 ; // returns the average price of product 2001
select Sum(pro_price*num) AS total returns the total value of the purchased product 2001
from tables
where pro_id =2001 ;
10) Group by group
Each supplier may provide multiple products, check the number of products provided by each supplier
select ven_id, COUNT(*) AS num_prods // Group count rows
form tables;
group by ven_id;
select
select province, countrycode, sum(popu) // country province city
from city
where countrycode = 'CHN'
group by countycode; //The original intention is to count the total population of each province in China, but here it is grouped by country, only the total population of China (where) is counted, and the grouping is not detailed enough
11) having filters each group individually
select province, countrycode, sum(popu) AS popu_of_province // country province city
from city
where countrycode = 'CHN'
group by countycode,province
having sum(popu)>200; count cities with a total population greater than 2 million in each province of China
Two subqueries (nested)
When using subqueries, it should be ensured that the select statement and the where statement have the same column information, and subqueries can also use calculated fields and statistical functions; pay attention to the use of fully qualified names
select cust_name, email // customer name, contact information
from customers // customer information table
Where cust_id IN ( select cust_id // cust_id 客户id
from orders // order image, including order number, customer id, order date, there is no actual order product information here
where order_num IN ( select order_num
from orderItems // The detailed information representation of the order saves the product id, price, quantity and other information of the order
where proc_id = 1002) ；
triple link
Table vendors ven_id (vendor id) vend_name address email
表products proc_id proc_name ven_id price
1 milk 1001 10.0
2 milk 1002 8.9
select vend_name, proc_name price
from vendors ,products // inner link
where vendors.ven_id = porducts.ven_id // Without the where clause, it will do the Cartesian product of the two tables, that is, combine any rows of the two tables, and output N*M rows
order by vend_name,proc_name;
select vend_name, proc_name price
from vendors INNER JOIN products // inner join
on vendors.ven_id = porducts.ven_id // Without the where clause, it will do the Cartesian product of the two tables, that is, combine any rows of the two tables, and output N*M rows
order by vend_name,proc_name;
A left OUTER JOIN B // Generate the complete set of table A, if there is a match in table B, there is a value, if there is no match, it is NULL
right OUTER JOIN