序言

show character set;
--查看mysql支持的字符集，部分輸出如下

show collation;
--查看mysql支持的collation

指定字符集與排序規(guī)則

服務(wù)器的字符集與排序規(guī)則

• character_set_server的默認(rèn)值為utf8mb4

  show variables like 'character_set_server';
  --輸出的Value列為utf8mb4
  

• 在my.ini中可以設(shè)置character_set_server的值：

  character-set-server=utf8;
  --在my.ini中添加上述行，重啟服務(wù)，即可設(shè)置character_set_server的值
  

• 在運(yùn)行過程中，可以更改character_set_server的值：

  set character_set_server = gbk;
  

• 如果未在create database語(yǔ)句中指定字符集和排序規(guī)則，則會(huì)將character_set_server和collation_server作為默認(rèn)值

• 若只指定了字符集沒有指定排序規(guī)則，則使用字符集對(duì)應(yīng)的默認(rèn)的排序規(guī)則

數(shù)據(jù)庫(kù)的字符集與排序規(guī)則

• 可以在create database和alter database時(shí)指定字符集和排序規(guī)則

  create database d character set gbk collate gbk_bin;
  alter database d character set gb2312 collate gb2312_bin;
  
  use d;
  SELECT @@character_set_database, @@collation_database;
  --此sql語(yǔ)句用于輸出當(dāng)前數(shù)據(jù)庫(kù)的字符集與排序規(guī)則
  

• 當(dāng)前數(shù)據(jù)庫(kù)的字符集和排序規(guī)則，可以通過查看character_set_database和collation_database來確定。若當(dāng)前沒有默認(rèn)數(shù)據(jù)庫(kù)，則character_set_database和collation_database與character_set_server和collation_server相同

  create database d0 character set gbk collate gbk_bin;
  create database d1 character set utf8 collate utf8_bin;
  
  use d0;
  show variables like 'character_set_database';
  --輸出的character_set_database值為gbk
  
  use d1;
  show variables like 'character_set_database';
  --輸出的character_set_database值為utf8
  

• 數(shù)據(jù)庫(kù)的字符集和排序規(guī)則會(huì)產(chǎn)生以下影響：

  • create table未指定字符集和排序規(guī)則時(shí)，則采用數(shù)據(jù)庫(kù)的字符集和排序規(guī)則

  • load data未指定字符集和排序規(guī)則，則使用數(shù)據(jù)庫(kù)的字符集來解釋文件中的信息

  • 創(chuàng)建的存儲(chǔ)過程和函數(shù)未指定字符集和排序規(guī)則，則使用數(shù)據(jù)庫(kù)的字符集和排序規(guī)則

    use d0; --create database d0 character set gbk collate gbk_bin;
    delimiter \
    
    create procedure p0(in str varchar(111))
    begin
            select charset(str);
    end\
    
    delimiter ;
    
    call p0("szn");
    --輸出:gbk
    

    use d1; --database d1 character set utf8 collate utf8_bin;
    delimiter \
    
    create procedure p1(in str varchar(111))
    begin
            select charset(str);
    end\
    
    delimiter ;
    
    call p1("szn");
    --輸出:utf8
    

    use d1; --database d1 character set utf8 collate utf8_bin;
    delimiter \
    
    create procedure p_set(in str varchar(111) character set gb2312)
    begin
            select charset(str);
    end\
    
    delimiter ;
    
    call p_set("szn");
    --輸出:gb2312
    

表的字符集與排序規(guī)則

• 可以在create table和alter table時(shí)指定表的字符集和排序規(guī)則

  create table t0(name varchar(1024)) character set gbk collate gbk_bin;
  alter table t0 character set utf8;
  

• 若未指定表的字符集和排序規(guī)則，則使用對(duì)應(yīng)數(shù)據(jù)庫(kù)的字符集和排序規(guī)則

• 若表中的列未指定字符集和排序規(guī)則，則對(duì)應(yīng)的列使用表的字符集和排序規(guī)則

列的字符集與排序規(guī)則

• 在create table和alter table時(shí)指定列的字符集和排序規(guī)則

  create table t2(name varchar(1024) character set gbk collate gbk_bin);
  alter table t2 modify column name varchar(1024) character set utf8 collate utf8_bin;
  

• 使用alter table將列的字符集進(jìn)行更改時(shí)，若字符集不兼容，則可能會(huì)丟失數(shù)據(jù)

字符串字面量的字符集與排序規(guī)則

• 字符串字面量的默認(rèn)字符集和排序規(guī)則由character_set_connection和collation_connection指定

• 字符串字面量可以設(shè)定字符集與排序規(guī)則：

  show variables like 'character_set_connection';
  --輸出值:utf8
  
  set @b = "szn";
  select charset(@b);
  --輸出值:utf8
  
  set @c = _gbk"哈" COLLATE gbk_bin;
  --字符串字面量前指定編碼方式，這種方式稱為介紹器(Character Set Introducers)
  
  select charset(@c);
  --輸出值:gbk
  

• 字符串字面量指定字符集并不會(huì)更改其值，解析器最終仍按照character_set_connection進(jìn)行處理

  set @c0 = _gbk"哈" COLLATE gbk_bin;
  set @c1 = _utf8"哈" COLLATE utf8_bin;
  select hex(@c0), hex(@c1);
  --兩個(gè)輸出均為"哈"的utf8編碼 E59388
  

Character Set Introducers

• 字符串字面量，十六進(jìn)制字面量，位字面量都有一個(gè)可選的字符集和排序規(guī)則設(shè)置，這稱為介紹器

• 介紹器告訴解析器后面字符串使用的字符集，但是不會(huì)改變字符串的值

  set @s0 = _utf8 0xE59388;
  set @s1 = _gbk 0xB9FE;
  set @s2 = _gbk X'B9FE';
  set @s3 = _utf8 b'111001011001001110001000';
  set @s4 = _utf8 0b111001011001001110001000;
  select @s0, @s1, @s2, @s3, @s4;
  --輸出:哈，哈，哈，哈, 哈
  

  SET @v1 = X'000D' | X'0BC0';
  --@V1是bigint類型
  
  SET @v2 = _binary X'000D' | X'0BC0';
  --@v2是字符串
  
  SELECT HEX(@v1), HEX(@v2);
  --輸出:BCD, 0BCD
  
  select @v1 + 1, @v2 + 1, "0BCD" + "1";.
  --輸出:3022, 1, 1
  

連接的字符集與排序規(guī)則

• 每個(gè)客戶端的連接都有特定且可更改的字符集和排序規(guī)則

• 客戶端發(fā)起sql查詢，至結(jié)果返回到客戶端：

  • character_set_client表明客戶端發(fā)送過來的sql語(yǔ)句的字符集
  • 服務(wù)器將接收到的sql語(yǔ)句的字符集轉(zhuǎn)換為character_set_connection
  • 服務(wù)器將sql的執(zhí)行結(jié)果的字符集轉(zhuǎn)換為character_set_results

• character_set_client不允許一些字符集設(shè)置：

  ucs2
  utf16
  utf16le
  utf32
  

• set character set會(huì)同時(shí)將character_set_client和character_set_results設(shè)定為給定值，并且將character_set_connection設(shè)置為character_set_database的值

  set character_set_client = ascii;
  set character_set_results = big5;
  set character_set_database = gb2312;
  set character_set_connection = gbk;
  
  set character set latin2;
  show variables like 'character%';
  

  Variable_name	Value
  character_set_client	latin2
  character_set_connection	gb2312
  character_set_database	gb2312
  character_set_results	latin2

• my.ini中添加default-character-set=latin2，重啟服務(wù)，將會(huì)同時(shí)更改character_set_client,character_set_connection,character_set_results為latin2

• 可以在登錄mysql時(shí):mysql -u root -p --default-character-set=latin1

• set names gb2312;
  --等效于下面三句代碼
  
  set character_set_client = gb2312;
  set character_set_results = gb2312;
  set character_set_connection = gb2312;
  

Unicode支持

? The Unicode Standard includes characters from the Basic Multilingual(使用多種語(yǔ)言的) Plane (BMP) and supplementary(追加的) characters that lie outside the BMP.

? BMP characters have these characteristics:

• Their code point values are between 0 and 65535 (or U+0000 and U+FFFF).
• They can be encoded in a variable-length encoding using 8, 16, or 24 bits (1 to 3 bytes).
• They can be encoded in a fixed-length encoding using 16 bits (2 bytes).
• They are sufficient(充分的) for almost all characters in major(主要的) languages.

? Supplementary characters lie outside the BMP:

• Their code point values are between U+10000 and U+10FFFF).
• 占用的空間比BMP內(nèi)的字符大，最多占用4字節(jié)

mysql支持的unicode：

• utf8當(dāng)前是utf8mb3的別稱，但是utf8mb3將在mysql將來的版本被刪除。有時(shí)候utf8會(huì)變成utf8mb4的別稱，所以為了避免模糊不清，請(qǐng)考慮使用utf8mb4

• 對(duì)于BMP字符，utf8mb4與utf8mb3具有相同編碼值，占用的空間也相同。即utf8mb4是utf8mb3的超集

• 對(duì)于補(bǔ)充字符，utf8mb4需要四個(gè)字節(jié)來存儲(chǔ)