因爲CHAR與VARCHAR只能用於存儲少量的字符串（1-65535個字符），所以當存儲大文本時，就需要用到TEXT或BLOB，其兩者的區別在於，TEXT只能保存字符串，而BLOB可以保存二進制數據（多媒體文件），見下表：
| 類型 | 可存儲字節 | 空間佔用 |
|
TINYTEXT/TINYBLOB |
256 bytes | 256bytes |
| TEXT/BLOB | 65,535 bytes | 64KB |
| MEDIUMTEXT/MEDIUMBLOB | 16,777,215 bytes | 16MB |
| LONGTEXT/LONGBLOB | 4,294,967,295 bytes | 4GB |
這兩種數據類型在執行刪除（尤其是大量刪除）時會呈現出一些問題：
刪除操作會在表中留下很多的“空洞”，在未來向這些“空洞”內插入數據時會有性能虧損，所以爲了提高性能，建議使用optimize table命令對錶進行壓縮,儘量的消除“空洞"
STEP1：創建數據庫
create database test;
STEP2：調用數據庫並創建表
create table test(id char(1),testchar text);
STEP3：創建test表
create table test(id char(1),testchar text);
STEP4：向test表中使用repeat函數插入大量字符串，插入1000個Hello
insert into test values(1,repeat('Hello',1000));
insert into test values(2,repeat('Hello',1000));
insert into test values(3,repeat('Hello',1000));
STEP5：反覆插入數據，語句意爲將整表數據插入表（複製）
insert into test select * from test;
Query OK, 3 rows affected (0.02 sec)
Records: 3 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 6 rows affected (0.00 sec)
Records: 6 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 12 rows affected (0.02 sec)
Records: 12 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 24 rows affected (0.00 sec)
Records: 24 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 48 rows affected (0.11 sec)
Records: 48 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 96 rows affected (0.13 sec)
Records: 96 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 192 rows affected (0.68 sec)
Records: 192 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 384 rows affected (0.04 sec)
Records: 384 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 768 rows affected (0.04 sec)
Records: 768 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 1536 rows affected (0.07 sec)
Records: 1536 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 3072 rows affected (0.12 sec)
Records: 3072 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 6144 rows affected (1.81 sec)
Records: 6144 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 12288 rows affected (3.15 sec)
Records: 12288 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 24576 rows affected (12.30 sec)
Records: 24576 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 49152 rows affected (22.52 sec)
Records: 49152 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 98304 rows affected (41.78 sec)
Records: 98304 Duplicates: 0 Warnings: 0
insert into test select * from test;
Query OK, 196608 rows affected (37.81 sec)
Records: 196608 Duplicates: 0 Warnings: 0
STEP6：查看物理文件大小，默認的mysql數據目錄在/var/lib/mysql數據庫與該目錄下的文件同名
cd /var/lib/mysql/test
du -sh test.ibd
2.1G test.ibd
STEP7：從數據庫中刪除所有ID=1的行
delete from test where id = 1;
Query OK, 131072 rows affected (33.21 sec)
STEP8：再進入數據庫目錄查看物理文件大小
du -sh test.ibd
2.1G test.ibd
STEP9：進入數據庫使用optimize table命令對錶進行碎片整理
optimize table test;
+-----------+----------+----------+-------------------------------------------------------------------+
| Table | Op | Msg_type | Msg_text |
+-----------+----------+----------+-------------------------------------------------------------------+
| test.test | optimize | note | Table does not support optimize, doing recreate + analyze instead |
| test.test | optimize | status | OK |
+-----------+----------+----------+-------------------------------------------------------------------+
STEP10：退出再次查看物理文件大小
du -sh test.ibd
1.6G test.ibd
可以從物理文件大小看出，表中的“空洞”已經被清空，這一特性與數據類型無關，是表本身呈現的問題