[sql] case expression

《SQL Advanced tutorial 》 MICK 1-1 CASE expression note

About... For creating tables for examples SQL Statements and sample code , Please refer to the following website .
http://www.ituring.com.cn/book/1813（ Please click on “ Download with book ” Download relevant materials in Chinese ）

---------------------------------------------------------------

explain ： The tests in the following notes are based on postgresql14

Command line connection local PSQL：  psql -U <username> -d <dbname> -h 127.0.0.1 -W

Catalog

 CASE Expression Overview

Two ways of writing ： Simple CASE expression 、 Search for CASE expression

Use CASE Note on expressions

Change the existing number into a new way and count

Use one SQL Statement for different condition statistics

stay UPDATE Conditional branching in a statement

---

CASE Expression Overview

Two ways of writing ： Simple CASE expression 、 Search for CASE expression

-- Simple CASE expression 
CASE sex
  WHEN '1' THEN ' male '
  WHEN '2' THEN ' Woman '
ELSE ' other ' END

-- Search for CASE expression 
CASE WHEN sex ='1' THEN ' male '
    WHEN sex ='2' THEN ' Woman '
ELSE ' other ' END

Use CASE Note on expressions

matters needing attention 1： Use WHEN Clause should pay attention to the exclusivity of conditions
Found to be true WHEN When clause ,CASE The true or false value judgment of the expression will be suspended , remainder WHEN Clause is ignored .

matters needing attention 2： Unify the data types returned by each branch

matters needing attention 3： Don't forget to write END

matters needing attention 4： Develop writing ELSE The habit of clauses
Don't write ELSE When clause ,CASE The result of the expression is NULL.

Change the existing number into a new way and count

PopTbl Create table command

/*  Convert the existing numbering method into a new one and make statistics  */
CREATE TABLE PopTbl
(pref_name VARCHAR(32) PRIMARY KEY,
 population INTEGER NOT NULL);

INSERT INTO PopTbl VALUES(' Tokushima ', 100);
INSERT INTO PopTbl VALUES(' Xiangchuan ', 200);
INSERT INTO PopTbl VALUES(' Ehime ', 150);
INSERT INTO PopTbl VALUES(' Gao Zhi ', 200);
INSERT INTO PopTbl VALUES(' Fukuoka, ', 300);
INSERT INTO PopTbl VALUES(' Saga ', 100);
INSERT INTO PopTbl VALUES(' Nagasaki ', 200);
INSERT INTO PopTbl VALUES(' Tokyo ', 400);
INSERT INTO PopTbl VALUES(' Herd horses ', 50);

PopTbl The contents are as follows

test=> select * from Poptbl;
 pref_name | population
-----------+------------
  Tokushima       |        100
  Xiangchuan       |        200
  Ehime       |        150
  Gao Zhi       |        200
  Fukuoka,       |        300
  Saga       |        100
  Nagasaki       |        200
  Tokyo       |        400
  Herd horses       |         50
(9 rows)

test=>

  First look at the following CASE Statement output result

test=> SELECT CASE pref_name
             WHEN ' Tokushima ' THEN ' Four countries '
             WHEN ' Xiangchuan ' THEN ' Four countries '
             WHEN ' Ehime ' THEN ' Four countries '
             WHEN ' Gao Zhi ' THEN ' Four countries '
             WHEN ' Fukuoka, ' THEN ' Kyushu '
             WHEN ' Saga ' THEN ' Kyushu '
             WHEN ' Nagasaki ' THEN ' Kyushu '
             ELSE ' other ' END AS district
  FROM PopTbl;
 district
----------
  Four countries 
  Four countries 
  Four countries 
  Four countries 
  Kyushu 
  Kyushu 
  Kyushu 
  other 
  other 
(9 rows)

test=> 

【 example 】 Convert the county number to the district number

test=> SELECT CASE pref_name
test->              WHEN ' Tokushima ' THEN ' Four countries '
test->              WHEN ' Xiangchuan ' THEN ' Four countries '
test->              WHEN ' Ehime ' THEN ' Four countries '
test->              WHEN ' Gao Zhi ' THEN ' Four countries '
test->              WHEN ' Fukuoka, ' THEN ' Kyushu '
test->              WHEN ' Saga ' THEN ' Kyushu '
test->              WHEN ' Nagasaki ' THEN ' Kyushu '
test->              ELSE ' other ' END AS district,
test->        SUM(population)
test->   FROM PopTbl
test->  GROUP BY CASE pref_name
test->              WHEN ' Tokushima ' THEN ' Four countries '
test->              WHEN ' Xiangchuan ' THEN ' Four countries '
test->              WHEN ' Ehime ' THEN ' Four countries '
test->              WHEN ' Gao Zhi ' THEN ' Four countries '
test->              WHEN ' Fukuoka, ' THEN ' Kyushu '
test->              WHEN ' Saga ' THEN ' Kyushu '
test->              WHEN ' Nagasaki ' THEN ' Kyushu '
test->              ELSE ' other ' END;
 district | sum
----------+-----
  Kyushu      | 600
  Four countries      | 650
  other      | 450
(3 rows)

test=>

GROUP BY The last paragraph can also be written district, as follows

test=> SELECT CASE pref_name
             WHEN ' Tokushima ' THEN ' Four countries '
             WHEN ' Xiangchuan ' THEN ' Four countries '
             WHEN ' Ehime ' THEN ' Four countries '
             WHEN ' Gao Zhi ' THEN ' Four countries '
             WHEN ' Fukuoka, ' THEN ' Kyushu '
             WHEN ' Saga ' THEN ' Kyushu '
             WHEN ' Nagasaki ' THEN ' Kyushu '
             ELSE ' other ' END AS district,
       SUM(population)
  FROM PopTbl
 GROUP BY district;
 district | sum
----------+-----
  Kyushu      | 600
  Four countries      | 650
  other      | 450
(3 rows)

test=>

explain ： there GROUP BY Clause uses exactly SELECT The nickname of the column defined in the clause district. But strictly speaking , This kind of writing is against the standard SQL Of the rules . because GROUP BY Clause ratio SELECT Statement first , So in GROUP BY Clause refers to in SELECT The nickname defined in clause is not allowed . in fact , stay Oracle、DB2、SQL Server When this method is used in the database, it will make mistakes . But there is also support for this SQL Statement database , For example, in PostgreSQL and MySQL in , This query statement can be executed smoothly .

Use one SQL Statement for different condition statistics

/*  Use one SQL Statement to make statistics of different conditions  */
SELECT pref_name,
       /*  Male population  */
       SUM( CASE WHEN sex = '1' THEN population ELSE 0 END) AS cnt_m,
       /*  Female population  */
       SUM( CASE WHEN sex = '2' THEN population ELSE 0 END) AS cnt_f
  FROM PopTbl2
 GROUP BY pref_name;

test->  GROUP BY pref_name;
 pref_name | cnt_m | cnt_f
-----------+-------+-------
  Nagasaki       |   125 |   125
  Gao Zhi       |   100 |   100
  Saga       |    20 |    80
  Tokyo       |   250 |   150
  Xiangchuan       |   100 |   100
  Ehime       |   100 |    50
  Tokushima       |    60 |    40
  Fukuoka,       |   100 |   200
(8 rows)

test=>

stay UPDATE Conditional branching in a statement

Suppose you need to update the data of this table according to the following conditions .
1. Yes, the current salary is 30 Employees above 10000 yen , A pay cut 10%.
2. Yes, the current salary is 25 More than 10000 yen and dissatisfied 28 Million yen employees , A raise 20%.
The updated data according to these requirements should be as shown in the following table .

Be careful ： You cannot use two in turn UPDATE Sentence completion , Because the second one UPDATE The statement will be based on the first UPDATE Result operation after statement update .

Salaries Create table statement

/*  Employee salary information table  */
CREATE TABLE Salaries
(name VARCHAR(32) PRIMARY KEY,
 salary INTEGER NOT NULL);

INSERT INTO Salaries VALUES(' Xiangtian ', 300000);
INSERT INTO Salaries VALUES(' Shenqi ', 270000);
INSERT INTO Salaries VALUES(' Wooden village ', 220000);
INSERT INTO Salaries VALUES(' saito  ', 290000);

  The update statement is as follows

/*  use CASE Write the correct update operation in the expression  */
UPDATE Salaries
   SET salary = CASE WHEN salary >= 300000
                     THEN salary * 0.9
                     WHEN salary >= 250000 AND salary < 280000
                     THEN salary * 1.2
                     ELSE salary END;