最近工作中，使用到了insert ignore into语法，感觉这个语法还是挺有用的，就记录下来做个总结。

insert ignore into : 忽略重复的记录，直接插入数据。

包括两种场景：

1、插入的数据是主键冲突时

insert ignore into会给出warnings，show warnings就可以看到提示主键冲突；并且本次插入无效。

[test]> create table tt(c1 int primary key, c2 varchar(50))engine = xx;
Query OK, 0 rows affected (0.21 sec)[test]> insert into tt values(1, "aaa"), (2, "bbb"), (3, "ccc");
Query OK, 3 rows affected (0.02 sec)
Records: 3  Duplicates: 0  Warnings: 0[test]> select * from tt;
+----+------+
| c1 | c2   |
+----+------+
|  1 | aaa  |
|  2 | bbb  |
|  3 | ccc  |
+----+------+
3 rows in set (0.01 sec)[test]> 
[test]> insert ignore into tt values(1, "aaa"), (2, "bbb"), (3, "ccc");
Query OK, 0 rows affected, 3 warnings (0.01 sec)
Records: 3  Duplicates: 3  Warnings: 3[test]> 
[test]> select * from tt;
+----+------+
| c1 | c2   |
+----+------+
|  1 | aaa  |
|  2 | bbb  |
|  3 | ccc  |
+----+------+
3 rows in set (0.00 sec)

使用insert ignore into语句时，如果主键冲突，只是提示"warnings"。

如果使用insert into语句时，如果主键冲突直接报错。

2、没有主键冲突时，直接插入数据

insert into 与 insert ignore into 都是直接插入数据

[test]> create table t2(c1 int, c2 varchar(50))engine = xxx;
Query OK, 0 rows affected (0.05 sec)[test]> insert into t2 values(1, "aaa"), (2, "bbb"), (3, "ccc");
Query OK, 3 rows affected (0.03 sec)
Records: 3  Duplicates: 0  Warnings: 0GreatDB Cluster[test]> select * from t2;
+------+------+
| c1   | c2   |
+------+------+
|    1 | aaa  |
|    2 | bbb  |
|    3 | ccc  |
+------+------+
3 rows in set (0.00 sec)[test]> insert into t2 values(1, "aaa"), (2, "bbb"), (3, "ccc");
Query OK, 3 rows affected (0.02 sec)
Records: 3  Duplicates: 0  Warnings: 0[test]> insert into t2 values(1, "aaa"), (2, "bbb"), (3, "ccc");
Query OK, 3 rows affected (0.02 sec)
Records: 3  Duplicates: 0  Warnings: 0[test]> select * from t2;
+------+------+
| c1   | c2   |
+------+------+
|    1 | aaa  |
|    2 | bbb  |
|    3 | ccc  |
|    1 | aaa  |
|    2 | bbb  |
|    3 | ccc  |
|    1 | aaa  |
|    2 | bbb  |
|    3 | ccc  |
+------+------+
9 rows in set (0.00 sec)[test]> insert ignore into t2 values(1, "aaa"), (2, "bbb"), (3, "ccc");
Query OK, 3 rows affected (0.03 sec)
Records: 3  Duplicates: 0  Warnings: 0

因此，insert ignore into主要是忽略重复的记录，直接插入数据。