is this bug?

[hanwooram]

Table name : HIST

HISTID	MANAGER_NAME	USER_NAME	USER_EMAIL	USERID
1	hong	chris	[email protected]	100
2	hong	evan	[email protected]	200
3	hong	jack	[email protected]	300
4	hong	mike	[email protected]	400
5	hong	ryon	[email protected]	500

<resultMap id="userResult" type="xxx.xxx.User">
        <result property="userId" column="USERID" />
        <result property="userName" column="USER_NAME" />
        <result property="email" column="USER_EMAIL" />
</resultMap>

<resultMap id="result" type="xxx.xxx.Hist">
        <result property="id" column="HISTID" />    
        <result property="managerName" column="MANAGER_NAME" />     
        <association property="user" javaType="xxx.xxx.User" resultMap="userResult" />
</resultMap>

<select id="findItems" parameterType="xxx.xxx.Hist" resultMap="result">
        SELECT MANAGER_NAME from HIST
</select>

Java:

List<Hist> list = histMapper.findItems(item);
System.out.println("LIST SIZE : " + list.size());

result:
DEBUG: xxxx.xxxx.HistMapper.findItems - <== Total: 5
LIST SIZE : 1 (why??)

But, modified query as below.

<select id="findItems" parameterType="xxx.xxx.Hist" resultMap="result">
                -- Added column HISTID
        SELECT HISTID, MANAGER_NAME from HIST
</select>

success!

or modified resultMap as below.

<resultMap id="userResult" type="xxx.xxx.User">
        <result property="userId" column="USERID" />
        <result property="userName" column="USER_NAME" />
        <result property="email" column="USER_EMAIL" />
    </resultMap>

    <resultMap id="result" type="xxx.xxx.Hist">
        <result property="id" column="HISTID" />    
        <result property="managerName" column="MANAGER_NAME" />     
                // removed association
        <-- association property="user" javaType="xxx.xxx.User" resultMap="userResult" -->
</resultMap>

success.

is this bug?

[mches]

I don't think this is a bug. For result maps with associations and collections MyBatis de-duplicates the result set according to the <id/> columns if present, otherwise every <result/> column. This is a very important feature when it comes to constructing a complex object model from a single result set. (i.e. using fetch joins to avoid the N+1 selects problem).

I'm assuming the column HISTID is the primary key of the HIST table. HISTID is not in your result set however, only MANAGER_NAME, which is apparently not unique. I would recommend to use a different <resultMap/> without associations or collections, or to include HISTID in your result set, by selecting it. For performance reasons, you should consider using <id/> instead of <result/> for the primary key columns.

[chanjarster]

agree with @mches , you need id or enough property mappings to uniquely identify each result row

[lidwt721]

agree with @mches , you need id or enough property mappings to uniquely identify each result row

agree with you,its useful to me too!