foreach一共有三种类型，分别为List,[](array),Map三种。
foreach属性
| 属性 | 描述 |
|---|---|
| item | 循环体中的具体对象。支持属性的点路径访问，如item.age,item.info.details。 具体说明：在list和数组中是其中的对象，在map中是value。 该参数为必选。 |
| collection | 要做foreach的对象，作为入参时，List<?>对象默认用list代替作为键，数组对象有array代替作为键，Map对象用map代替作为键。 当然在作为入参时可以使用@Param("keyName")来设置键，设置keyName后，list,array,map将会失效。 除了入参这种情况外，还有一种作为参数对象的某个字段的时候。举个例子： 如果User有属性List ids。入参是User对象，那么这个collection = "ids" 如果User有属性Ids ids;其中Ids是个对象，Ids有个属性List id;入参是User对象，那么collection = "ids.id" 上面只是举例，具体collection等于什么，就看你想对那个元素做循环。 该参数为必选。 |
| separator | 元素之间的分隔符，例如在in()的时候，separator=","会自动在元素中间用“,“隔开，避免手动输入逗号导致sql错误，如in(1,2,)这样。该参数可选。 |
| open | foreach代码的开始符号，一般是(和close=")"合用。常用在in(),values()时。该参数可选。 |
| close | foreach代码的关闭符号，一般是)和open="("合用。常用在in(),values()时。该参数可选。 |
| index | 在list和数组中,index是元素的序号，在map中，index是元素的key，该参数可选。 |
示例一：
<select id="countByUserList" resultType="_int" parameterType="list">
select count(*) from users
<where>
id in
<foreach item="item" collection="list" separator="," open="(" close=")" index="">
#{item.id, jdbcType=NUMERIC}
</foreach>
</where>
</select>
注：select count(*) from users WHERE id in ( ? , ? )
示例二：
<select id="selectStorageProductInventroy" parameterType="java.util.Map" resultType="StorageProductInventroyModel">
select op.inventroy_operation_id ,
op.inventroy_id ,
ty.name as operation_type,
DATE_FORMAT(op.insert_time,'%Y/%m/%d') as insert_time,
em.name as employee,
sm1.name as from_store,
st1.name as to_store,
sm.name as from_storage,
st.name as to_storage,
hi.number,
pr.name as product_name,
pr.category_id as categoryString,
pr.product_code,
hi.amount,
hi.back_amount,
hi.unit,
re.description as operation_reason,
hi.description
from inventroy_history hi left join inventroy_operation op
on hi.inventroy_operation_id = op.inventroy_operation_id
left join product pr
on hi.product_id =pr.product_id
left join operation_type ty
on op.operation_type_id = ty.operation_type_id
left join employee em
on op.employee_id = em.employee_id
left join (select inventroy_operation.to_storage_id as to_storage_id,storage.name as name
from inventroy_operation , storage
where inventroy_operation.to_storage_id = storage.storage_id) st
on op.to_storage_id = st.to_storage_id
left join (select inventroy_operation.from_storage_id as from_storage_id,storage.name as name
from inventroy_operation , storage
where inventroy_operation.from_storage_id = storage.storage_id) sm
on op.from_storage_id = sm.from_storage_id
left join (select inventroy_operation.to_store_id as to_store_id,store.name as name
from inventroy_operation , store
where inventroy_operation.to_store_id = store.store_id) st1
on op.to_store_id = st1.to_store_id
left join (select inventroy_operation.from_store_id as from_store_id,store.name as name
from inventroy_operation , store
where inventroy_operation.from_store_id = store.store_id) sm1
on op.from_store_id = sm1.from_store_id
left join operation_reason re
on hi.operation_reason_id = re.operation_reason_id
WHERE 1=1
<if test="operation_type_id != null and operation_type_id.length != 0">
and
<foreach collection="operation_type_id" item="operationTypeId" index="index" open="(" separator="or" close=")">
op.operation_type_id = #{operationTypeId}
</foreach>
</if>
<if test="storage_id != 0 and storage_id != null">
and (op.from_storage_id = #{storage_id} or op.to_storage_id = #{storage_id})
</if>
<if test="category_id != 0 and category_id != null">
and pr.category_id like concat('%','(',#{category_id},')','%')
</if>
<if test="product_code != null">
and pr.product_code like concat('%',#{product_code},'%')
</if>
<if test="product_name != null">
and pr.name like concat('%',#{product_name},'%')
</if>
and date(op.insert_time) between #{start_time} and #{end_time}
group by hi.inventroy_history_id
</select>
注：and （op.operation_type_id = ？ or op.operation_type_id = ？）
示例三：数组
public void testQuery() {
ColInfoDao dao=(ColInfoDao)ctx.getBean("colInfoDao");
Map map = new HashMap();
map.put("userId", "tom");
map.put("password", "123");
String[] a = { "20000001", "20000002" };
map.put("classIds", Arrays.asList(a));
Object password = dao.query(map);
System.out.println("password:" + password);
Assert.assertEquals("123", password);
}
XML：（感觉不适合mybatis，可以使用在ibatis中，iBatis 2.x 和 MyBatis 3.0.x）
<select id="queryPasswordByUserId" parameterClass="java.util.Map" resultClass="java.lang.String">
<![CDATA[
select PASSWORD as password from T_S_P_USER
]]>
<dynamic prepend="where">
<isNotEmpty prepend="AND" property="userId">
USER_ID=#userId#
</isNotEmpty>
<isNotEmpty prepend="AND" property="password">
PASSWORD=#password#
</isNotEmpty>
<isNotEmpty prepend="AND" property="classIds">
<iterate property="classIds" open="(" conjunction="OR" close=")">
CLASS_ID = #classIds[]#
</iterate>
</isNotEmpty>
</dynamic>
</select>
示例四：Map
map和List,array相比，map是用K,V存储的，在foreach中，使用map时，index属性值为map中的Key的值。
因为map中的Key不同于list,array中的索引，所以会有更丰富的用法。
<insert id="ins_string_string">
insert into string_string (key, value) values
<foreach item="item" index="key" collection="map"
open="" separator="," close="">(#{key}, #{item})</foreach>
</insert>
可以看到这个例子相当简单，表中需要两个值，正好和K,V对应，因而map中的一个K,V就对应一条数据，如果map中有多个K,V，就会保存多个结果。
如果map中有两对K,V，那么执行SQL如下：
DEBUG [main] - ==> Preparing: insert into string_string (key, value) values (?, ?) , (?, ?)
DEBUG [main] - ==> Parameters: key 1(String), value 1(String), key 2(String), value 2(String)
DEBUG [main] - <== Updates: 2
下面再看一个select的例子：
<select id="sel_key_cols" resultType="int">
select count(*) from key_cols where
<foreach item="item" index="key" collection="map"
open="" separator="AND" close="">${key} = #{item}</foreach>
</select>
可以看到这里用key=value来作为查询条件，对于动态的查询，这种处理方式可以借鉴。一定要注意到$和#的区别，$的参数直接输出，#的参数会被替换为?，然后传入参数值执行。
DEBUG [main] - ==> Preparing: select count(*) from key_cols where col_a = ? AND col_b = ?
DEBUG [main] - ==> Parameters: 22(Integer), 222(Integer)
DEBUG [main] - <== Total: 1
最后，如果不考虑元素的顺序和map中Key，map和list,array可以拥有一样的效果，都是存储了多个值，然后循环读取出来。