一、提出疑问
IN 操作语句：
SELECT column_name(s) FROM table_name WHERE column_name IN
(value1,value2,…)；
疑问：column_name的数据类型是否要与value1、value2数据类型一致？？in括号里面的value1、value2…是否要加单引号，有什么区别？？
二、情况分析
in的单字段匹配
举例：如果column_name类型是bigInt、double或者string，in()括号里面的值加不加单引号，都可以匹配成功。
> create table test_table1
(action_id bigint,
value string);
> insert overwrite table test_table1 values ('2003140','12'),('2003160','23'),('14027','34');
> select * from test_table1;
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
| 14027 | 34 |
+------------------------+--------------------+
## action_id是bigint类型，in()括号里面的值加单引号。
> select * from test_table1 where action_id in ('2003140','2003160','14027');
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
| 14027 | 34 |
+------------------------+--------------------+
## action_id是bigint类型，in()括号里面的值不加单引号。
> select * from test_table1 where action_id in (2003140,2003160,14027);
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
| 14027 | 34 |
+------------------------+--------------------+
> create table test_table2
(action_id double,
value string);
> insert overwrite table test_table2 values ('2003140','12'),('2003160','23'),('14027','34');
> select * from test_table2;
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
| 2003160.0 | 23 |
| 14027.0 | 34 |
+------------------------+--------------------+
## action_id是double类型，in()括号里面的值加单引号。
> select * from test_table2 where action_id in ('2003140','2003160','14027');
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
| 2003160.0 | 23 |
| 14027.0 | 34 |
+------------------------+--------------------+
## action_id是double类型，in()括号里面的值不加单引号。
> select * from test_table2 where action_id in (2003140,2003160,14027);
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
| 2003160.0 | 23 |
| 14027.0 | 34 |
+------------------------+--------------------+
> create table test_table3
(action_id string,
value string);
> insert overwrite table test_table3 values ('2003140','12'),('2003160','23'),('14027','34');
> select * from test_table3;
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
| 14027 | 34 |
+------------------------+--------------------+
## action_id是string类型，in()括号里面的值加单引号。
> select * from test_table3 where action_id in ('2003140','2003160','14027');
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
| 14027 | 34 |
+------------------------+--------------------+
## action_id是string类型，in()括号里面的值不加单引号。
> select * from test_table3 where action_id in (2003140,2003160,14027);
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
| 14027 | 34 |
+------------------------+--------------------+
总结：单字段匹配，in括号里面的值加不加单引号都可以匹配上。
in多字段匹配
多个字段匹配一种情况
举例1：action_id是bigint类型，value是string类型
# 保持原有类型
> select * from test_table1 where (action_id,value) in ((2003140,12));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table1 where (action_id,value) in (('2003140','12'));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
+------------------------+--------------------+
## 两者同时为string
> select * from test_table1 where (cast(action_id as string),value) in ((2003140,12));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table1 where (cast(action_id as string),value) in (('2003140','12'));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
+------------------------+--------------------+
## 两者同时为bigint
> select * from test_table1 where (action_id ,cast(value as bigint)) in ((2003140,12));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table1 where (action_id ,cast(value as bigint)) in (('2003140','12'));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
+------------------------+--------------------+
结论：无论是action_id与value是分别为bigint，string或者同时为bigint或者同时为string类型，in括号里面必须加单引号才能匹配成功。不考虑in括号里面一个带引号，一个不带引号。
举例2：action_id是double类型，value是string类型
# 保持原有类型
> select * from test_table2 where (action_id,value) in ((2003140,12));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id,value) in ((2003140.0,12));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id,value) in (('2003140','12'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id,value) in (('2003140.0','12'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
+------------------------+--------------------+
# 两者同时为string
> select * from test_table2 where (cast(action_id as string),value) in ((2003140,12));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (cast(action_id as string),value) in ((2003140.0,12));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (cast(action_id as string),value) in (('2003140','12'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (cast(action_id as string),value) in (('2003140.0','12'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
+------------------------+--------------------+
# 两者同时为double
> select * from test_table2 where (action_id ,cast(value as double)) in ((2003140,12));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id ,cast(value as double)) in ((2003140.0,12));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id ,cast(value as double)) in (('2003140','12'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id ,cast(value as double)) in (('2003140.0','12'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
结论：
1、注意：double类型字段显示为2003140.0，因些(2003140,12)与(‘2003140’,‘12’)都无法匹配成功，要带单引号的(‘2003140.0’,‘12’)才能匹配成功，在单个字段匹配里面无所谓"2003140"与"2003140.0"。
2、无论是action_id与value是分别为double，string或者同时为string类型，in括号里面必须加单引号才能匹配成功，同时为double类型，无论加不加引号都匹配不成功。不考虑in括号里面一个带引号，一个不带引号。
举例3：action_id是string类型，value是string类型
> select * from test_table3 where (action_id,value) in ((2003140,12));
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table3 where (action_id,value) in (('2003140','12'));
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
| 2003140 | 12 |
+------------------------+--------------------+
结论：action_id与value同时为string类型，in括号里面必须加单引号才能匹配成功。不考虑in括号里面一个带引号，一个不带引号。
多个字段匹配多种情况
举例1：action_id是bigint类型，value是string类型
## 保持原有类型
> select * from test_table1 where (action_id,value) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table1 where (action_id,value) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
+------------------------+--------------------+
## 两者转换成string
> select * from test_table1 where (cast(action_id as string),value) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table1 where (cast(action_id as string),value) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
+------------------------+--------------------+
## 两者转换成bigint
> select * from test_table1 where (action_id ,cast(value as bigint)) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table1 where (action_id ,cast(value as bigint)) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table1.action_id | test_table1.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
+------------------------+--------------------+
结论：无论是action_id与value是分别为bigint，string或者同时为bigint或者同时为string类型，in括号里面必须加单引号才能匹配成功。不考虑in括号里面一个带引号，一个不带引号。
举例2：action_id是double类型，value是string类型
## 保持原有类型
> select * from test_table2 where (action_id,value) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id,value) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id,value) in ((2003140.0,12),(2003160.0,23));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id,value) in (('2003140.0','12'),('2003160.0','23'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
| 2003160.0 | 23 |
+------------------------+--------------------+
## 两者转换成string
> select * from test_table2 where (cast(action_id as string),value) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (cast(action_id as string),value) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (cast(action_id as string),value) in ((2003140.0,12),(2003160.0,23));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (cast(action_id as string),value) in (('2003140.0','12'),('2003160.0','23'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
| 2003140.0 | 12 |
| 2003160.0 | 23 |
+------------------------+--------------------+
## 两者转换成double
> select * from test_table2 where (action_id ,cast(value as double)) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id ,cast(value as double)) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id ,cast(value as double)) in ((2003140.0,12),(2003160.0,23));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table2 where (action_id ,cast(value as double)) in (('2003140.0','12'),('2003160.0','23'));
+------------------------+--------------------+
| test_table2.action_id | test_table2.value |
+------------------------+--------------------+
+------------------------+--------------------+
结论：
1、注意：double类型字段显示为"2003140.0"，“2003160.0”，因些(2003140,12)与(‘2003140’,‘12’),(2003160,23)与(‘2003160’,‘23’)都无法匹配成功，要带单引号的(‘2003140.0’,‘12’)与(‘2003160.0’,‘23’)才能匹配成功，在单个字段匹配里面无所谓"2003140"与"2003140.0"。
2、无论是action_id与value是分别为double，string或者同时为string类型，in括号里面必须加单引号才能匹配成功，同时为double类型，无论加不加引号都匹配不成功。不考虑in括号里面一个带引号，一个不带引号。
举例3：action_id是string类型，value是string类型
> select * from test_table3 where (action_id,value) in ((2003140,12),(2003160,23));
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
+------------------------+--------------------+
> select * from test_table3 where (action_id,value) in (('2003140','12'),('2003160','23'));
+------------------------+--------------------+
| test_table3.action_id | test_table3.value |
+------------------------+--------------------+
| 2003140 | 12 |
| 2003160 | 23 |
+------------------------+--------------------+
结论：action_id与value同时为string类型，in括号里面必须加单引号才能匹配成功。不考虑in括号里面一个带引号，一个不带引号。
三、总结
1、单个字段匹配，与数据类型无关，in括号里面字段加不加单引号进行匹配都行，都可以匹配成功。
2、多字段匹配，in括号里面字段必须要加单引号进行匹配，否则匹配不上，为空。
3、还要特别注意double类型，double类型就算两个字段都转换成string类型，匹配in括号里面加单引号字段也无法匹配；但其他数据类型，如bigint，转换成string格式，匹配in括号里面加单引号字段可以匹配成功。