mybatis中orderBy(排序字段)和sort(排序方式)引起的bug及解决
引言
记录一个mybatis实现动态字段的排序和动态的升降序问题
实现效果如下:
问题叙述
在这里无论使用postman是否传递sort的值， 都不生效，
postman
执行的sql日志
==> preparing: select sum(acd.read_view_count) as read_view_count,sum(acd.read_person_count) as read_person_count, sum(acd.like_count) as like_count,sum(acd.comment_count) as comment_count, sum(acd.collect_count) as collect_count,sum(acd.transmit_count) as transmit_count, avg(acd.avg_stay_time) as avg_stay_time,avg(acd.visit_count) as visit_count, acd.article_id, acd.article_title from article_content_data acd left join cplatform cp on cp.cid = acd.platform_id where 1=1 group by acd.article_id desc order by transmit_count
==> parameters:
<== columns: read_view_count, read_person_count, like_count, comment_count, collect_count, transmit_count, avg_stay_time, visit_count, article_id, article_title
<== row: 900, 900, 960, 390, 12592, 423, 1548.3333, 1.0000, 3, 这是一个正经的文章标题3
<== row: 1500, 1500, 1560, 390, 1497, 423, 5294.6667, 1.0000, 5, 这是一个正经的文章标题5
<== row: 400, 440, 480, 520, 526, 546, 419.0000, 1.0000, 1, 震惊, 百分之99的人在喝了睡之后, 100年之内会死
<== row: 800, 800, 880, 520, 2587, 777, 426.2500, 1.0000, 2, 这是一个正经的文章标题2
<== row: 1200, 1200, 1260, 390, 386, 816, 121.0000, 1.0000, 4, 这是一个正经的文章标题4
<== total: 5
**很明显, order by之后只拼了排序字段, 而没有拼上排序方式 **
下面尝试采用第二种方式
将排序sql语句替换为以下这个
order by
${orderby}
<choose>
<when test="sort!=null">
${sort}
</when>
<otherwise>
asc
</otherwise>
</choose>
完整代码如下
<select id="listarticlecontentdatadto" resulttype="com.etouch.pojo.dto.data.articlecontentdatadto">
select
sum(acd.read_view_count) as read_view_count,sum(acd.read_person_count) as read_person_count,
sum(acd.like_count) as like_count,sum(acd.comment_count) as comment_count,
sum(acd.collect_count) as collect_count,sum(acd.transmit_count) as transmit_count,
avg(acd.avg_stay_time) as avg_stay_time,avg(acd.visit_count) as visit_count,
acd.article_id, acd.article_title
from article_content_data acd
left join cplatform cp on cp.cid = acd.platform_id
<include refid="searcharticledata"/>
group by acd.article_id
order by
${orderby}
<choose>
<when test="sort!=null">
${sort}
</when>
<otherwise>
asc
</otherwise>
</choose>
</select>
执行查询,结果如下
==> preparing: select sum(acd.read_view_count) as read_view_count, sum(acd.read_person_count) as read_person_count, sum(acd.like_count) as like_count, sum(acd.comment_count) as comment_count, sum(acd.collect_count) as collect_count, sum(acd.transmit_count) as transmit_count, avg(acd.avg_stay_time) as avg_stay_time, avg(acd.visit_count) as visit_count, acd.article_id, acd.article_title from article_content_data acd left join cplatform cp on cp.cid = acd.platform_id where 1 = 1 group by acd.article_id order by transmit_count
==> parameters:
<== columns: read_view_count, read_person_count, like_count, comment_count, collect_count, transmit_count, avg_stay_time, visit_count, article_id, article_title
<== row: 900, 900, 960, 390, 12592, 423, 1548.3333, 1.0000, 3, 这是一个正经的文章标题3
<== row: 1500, 1500, 1560, 390, 1497, 423, 5294.6667, 1.0000, 5, 这是一个正经的文章标题5
<== row: 400, 440, 480, 520, 526, 546, 419.0000, 1.0000, 1, 震惊, 百分之99的人在喝了睡之后, 100年之内会死
<== row: 800, 800, 880, 520, 2587, 777, 426.2500, 1.0000, 2, 这是一个正经的文章标题2
<== row: 1200, 1200, 1260, 390, 386, 816, 121.0000, 1.0000, 4, 这是一个正经的文章标题4
<== total: 5
同样是没有实现排序,此时应该考虑,mybatis是否无法同时接受orderby和sort都是变量的情况, 带着这个疑问, 将order by的排序字段设置为固定值, sort设置为前端传递过来的值, 代码如下
order by
acd.like_count
<choose>
<when test="sort!=null">
${sort}
</when>
<otherwise>
asc
</otherwise>
</choose>
执行sql日志如下
==> preparing: select sum(acd.read_view_count) as read_view_count,sum(acd.read_person_count) as read_person_count, sum(acd.like_count) as like_count,sum(acd.comment_count) as comment_count, sum(acd.collect_count) as collect_count,sum(acd.transmit_count) as transmit_count, avg(acd.avg_stay_time) as avg_stay_time,avg(acd.visit_count) as visit_count, acd.article_id, acd.article_title from article_content_data acd left join cplatform cp on cp.cid = acd.platform_id where 1=1 group by acd.article_id order by acd.like_count asc
==> parameters:
<== columns: read_view_count, read_person_count, like_count, comment_count, collect_count, transmit_count, avg_stay_time, visit_count, article_id, article_title
<== row: 400, 440, 480, 520, 526, 546, 419.0000, 1.0000, 1, 震惊, 百分之99的人在喝了睡之后, 100年之内会死
<== row: 800, 800, 880, 520, 2587, 777, 426.2500, 1.0000, 2, 这是一个正经的文章标题2
<== row: 900, 900, 960, 390, 12592, 423, 1548.3333, 1.0000, 3, 这是一个正经的文章标题3
<== row: 1200, 1200, 1260, 390, 386, 816, 121.0000, 1.0000, 4, 这是一个正经的文章标题4
<== row: 1500, 1500, 1560, 390, 1497, 423, 5294.6667, 1.0000, 5, 这是一个正经的文章标题5
<== total: 5
closing non transactional sqlsession [org.apache.ibatis.session.defaults.defaultsqlsession@32112885]
这里发现实现了升降序, 但是这样不能满足原型上的要求, 动态的字段排序和升降序
这里不妨换种方式思考, 假如将排序字段和排序方式拼的一起, 存放到一个字段里呢? 带着这个猜想, 进行第三种方式
第三种方式
代码
<choose>
<when test="orderby!=null and orderby!=''">
order by ${orderby}
</when>
<otherwise>
order by acd.create_time asc
</otherwise>
</choose>
3.1 首先是什么都不传
直接sql结果如下
==> preparing: select sum(acd.read_view_count) as read_view_count,sum(acd.read_person_count) as read_person_count, sum(acd.like_count) as like_count,sum(acd.comment_count) as comment_count, sum(acd.collect_count) as collect_count,sum(acd.transmit_count) as transmit_count, avg(acd.avg_stay_time) as avg_stay_time,avg(acd.visit_count) as visit_count, acd.article_id, acd.article_title from article_content_data acd left join cplatform cp on cp.cid = acd.platform_id where 1=1 group by acd.article_id order by acd.create_time asc
==> parameters:
<== columns: read_view_count, read_person_count, like_count, comment_count, collect_count, transmit_count, avg_stay_time, visit_count, article_id, article_title
<== row: 400, 440, 480, 520, 526, 546, 419.0000, 1.0000, 1, 震惊, 百分之99的人在喝了睡之后, 100年之内会死
<== row: 800, 800, 880, 520, 2587, 777, 426.2500, 1.0000, 2, 这是一个正经的文章标题2
<== row: 900, 900, 960, 390, 12592, 423, 1548.3333, 1.0000, 3, 这是一个正经的文章标题3
<== row: 1200, 1200, 1260, 390, 386, 816, 121.0000, 1.0000, 4, 这是一个正经的文章标题4
<== row: 1500, 1500, 1560, 390, 1497, 423, 5294.6667, 1.0000, 5, 这是一个正经的文章标题5
<== total: 5
默认采用创建日期排序, 避免前端不传排序字段导致报错
3.2 传入排序字段oderby
这里将排序字段和排序方式用空格隔开, 存入orderby中
查看sql执行结果
==> preparing: select sum(acd.read_view_count) as read_view_count, sum(acd.read_person_count) as read_person_count, sum(acd.like_count) as like_count, sum(acd.comment_count) as comment_count, sum(acd.collect_count) as collect_count, sum(acd.transmit_count) as transmit_count, avg(acd.avg_stay_time) as avg_stay_time, avg(acd.visit_count) as visit_count, acd.article_id, acd.article_title from article_content_data acd left join cplatform cp on cp.cid = acd.platform_id where 1 = 1 group by acd.article_id order by like_count desc
==> parameters:
<== columns: read_view_count, read_person_count, like_count, comment_count, collect_count, transmit_count, avg_stay_time, visit_count, article_id, article_title
<== row: 1500, 1500, 1560, 390, 1497, 423, 5294.6667, 1.0000, 5, 这是一个正经的文章标题5
<== row: 1200, 1200, 1260, 390, 386, 816, 121.0000, 1.0000, 4, 这是一个正经的文章标题4
<== row: 900, 900, 960, 390, 12592, 423, 1548.3333, 1.0000, 3, 这是一个正经的文章标题3
<== row: 800, 800, 880, 520, 2587, 777, 426.2500, 1.0000, 2, 这是一个正经的文章标题2
<== row: 400, 440, 480, 520, 526, 546, 419.0000, 1.0000, 1, 震惊, 百分之99的人在喝了睡之后, 100年之内会死
<== total: 5
很明显, 实现了动态排序, 大功告成, 接下来我会继续研究,到底是为什么myabtis不能同时接受orderby和sort同时为变量的情况, 欢迎评论区补充
ps：order by或者group by后面不可以使用#{变量}的方式, 因为mybatis会把#{变量}进行预编译, 这也是为了防止sql注入, 也就是编译成了order by ‘字段名’ , 也就会导致排序失败, 这里只能采用${变量}的方式, ${}中的内容会不经过编译直接拼到order by后面, 但是不能避免sql注入, 可以再后台通过限制order by变量长度的方式, 来避免sql注入, 是麻烦了点, 但是鱼和熊掌不可兼得
以上为个人经验，希望能给大家一个参考，也希望大家多多支持。