Mysql varchar Size length problem introduction _mysql

Version 4.0 below, varchar (20), refers to 20 bytes, if stored UTF8 Chinese characters, can only save 6 (each Chinese character 3 bytes)
5.0 version above, varchar (20), refers to 20 characters, whether stored in numbers, letters or UTF8 characters (3 bytes per character), can store 20, the maximum size is 65532 bytes
The maximum number of Mysql4 is only 20 bytes, but the storage size varies according to the encoding Mysql5.

1. Restriction rules

The limits of a field have the following rules when the field is defined:

a) storage limits

The varchar field is to store the actual content separately from the clustered index, where the content begins with 1 to 2 bytes representing the actual length (2 bytes in length over 255), so the maximum length cannot exceed 65535.

b Coding length Limit

If the character type is GBK, each character is up to 2 bytes and the maximum length cannot exceed 32766;

If the character type is UTF8, each character is up to 3 bytes, and the maximum length cannot exceed 21845.

If the definition exceeds the above limit, the varchar field is forcibly converted to the text type and produces warning.

c) The length of the limit

The length of a row definition is the limit of the varchar length in the actual application. MySQL requires a row to have a defined length of not more than 65535. If the defined table length exceeds this value, the prompt

ERROR 1118 (42000): Row size too large. The maximum row size for the used table type, not counting BLOBs, is 65535. You are have to change some columns to TEXT or BLOBs.

2. Calculation Example

Give two examples to illustrate the actual length of the calculation.

A If a table has only one varchar type, as defined as

CREATE table t4 (c varchar (N)) CHARSET=GBK;

Then the maximum value for n here is (65535-1-2)/2= 32766.

The reason for minus 1 is that the actual row is stored starting with the second byte ';

The reason for minus 2 is that the length of the varchar head is 2 bytes;

The reason for the addition of 2 is that the character encoding is GBK.

b If a table is defined as

CREATE table t4 (c int, C2 char (), C3 varchar (N)) Charset=utf8;

Then the maximum value for n here is (65535-1-2-4-30*3)/3=21812

Minus 1 and minus 2 are the same as the previous example;

The reason for minus 4 is that C of type int is 4 bytes;

The reason for reducing the 30*3 is that char (30) occupies 90 bytes and the encoding is UTF8.

If the varchar exceeds the above B rule and is strongly converted to the text type, then each field occupies a defined length of 11 bytes, of course this is not "varchar" anymore.