SQL 语法速览与备忘清单

SQL 是 Structrued Query Language 的缩写,即结构化查询语言。它是负责与 ANSI(美国国家标准学会)维护的数据库交互的标准。作为关系数据库的标准语言,它已被众多商用 DBMS 产品所采用,使得它已成为关系数据库领域中一个主流语言,不仅包含数据查询功能,还包括插入、删除、更新和数据定义功能。最为重要的 SQL92 版本的详细标准可以查看这里,或者在 Wiki 上查看 SQL 标准的变化。

-- Comments start with two hyphens. End each command with a semicolon.
-- SQL is not case-sensitive about keywords. The sample commands here
-- follow the convention of spelling them in upper-case because it makes
-- it easier to distinguish them from database, table, and column names.
-- Create and delete a database. Database and table names are case-sensitive.
DROP DATABASE someDatabase;
-- List available databases.
-- Use a particular existing database.
USE employees;
-- Select all rows and columns from the current database's departments table.
-- Default activity is for the interpreter to scroll the results on your screen.
SELECT * FROM departments;
-- Retrieve all rows from the departments table,
-- but only the dept_no and dept_name columns.
-- Splitting up commands across lines is OK.
SELECT dept_no,
dept_name FROM departments;
-- Retrieve all departments columns, but just 5 rows.
SELECT * FROM departments LIMIT 5;
-- Retrieve dept_name column values from the departments
-- table where the dept_name value has the substring 'en'.
SELECT dept_name FROM departments WHERE dept_name LIKE '%en%';
-- Retrieve all columns from the departments table where the dept_name
-- column starts with an 'S' and has exactly 4 characters after it.
SELECT * FROM departments WHERE dept_name LIKE 'S____';
-- Select title values from the titles table but don't show duplicates.
-- Same as above, but sorted (case-sensitive) by the title values.
-- Show the number of rows in the departments table.
SELECT COUNT(*) FROM departments;
-- Show the number of rows in the departments table that
-- have 'en' as a substring of the dept_name value.
SELECT COUNT(*) FROM departments WHERE dept_name LIKE '%en%';
-- A JOIN of information from multiple tables: the titles table shows
-- who had what job titles, by their employee numbers, from what
-- date to what date. Retrieve this information, but instead of the
-- employee number, use the employee number as a cross-reference to
-- the employees table to get each employee's first and last name
-- instead. (And only get 10 rows.)
SELECT employees.first_name, employees.last_name,
titles.title, titles.from_date, titles.to_date
FROM titles INNER JOIN employees ON
employees.emp_no = titles.emp_no LIMIT 10;
-- List all the tables in all the databases. Implementations typically provide
-- their own shortcut command to do this with the database currently in use.
-- Sub Query
FROM section
WHERE semester = ‘Fall’ AND year= 2009 AND course_id IN (
SELECT course_id
FROM section
WHERE semester = ‘Spring’ AND year= 2010
-- Create a table called tablename1, with the two columns shown, for
-- the database currently in use. Lots of other options are available
-- for how you specify the columns, such as their datatypes.
CREATE TABLE <table_name1> (
<col_name1> <col_type1>,
<col_name2> <col_type2>,
<col_name3> <col_type3>
PRIMARY KEY (<col_name1>),
FOREIGN KEY (<col_name2>) REFERENCES <table_name2>(<col_name2>)
CREATE TABLE tablename1 (fname VARCHAR(20), lname VARCHAR(20));
-- Insert a row of data into the table tablename1. This assumes that the
-- table has been defined to accept these values as appropriate for it.
INSERT INTO tablename1 VALUES('Richard','Mutt');
-- In tablename1, change the fname value to 'John'
-- for all rows that have an lname value of 'Mutt'.
UPDATE tablename1 SET fname='John' WHERE lname='Mutt';
-- Delete rows from the tablename1 table
-- where the lname value begins with 'M'.
DELETE FROM tablename1 WHERE lname like 'M%';
-- Delete all rows from the tablename1 table, leaving the empty table.
DELETE FROM tablename1;
-- Remove the entire tablename1 table.
DROP TABLE tablename1;