我有以下数据库模型
create table Diary (id bigint NOT NULL AUTO_INCREMENT,
creationDate TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
name varchar(255) not null,
description text,
viewtype varchar(255) not null,
member bigint,
primary key (id),
foreign key (member) references Member(id));
create table Page (id bigint NOT NULL AUTO_INCREMENT,
creationDate TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
viewtype varchar(255) not null,
diary bigint,
member bigint,
primary key (id),
foreign key (diary) references Diary(id),
foreign key (member) references Member(id));
create table Comment (id bigint NOT NULL AUTO_INCREMENT,
postingDate TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
comment text not null,
page bigint,
member bigint,
primary key (id),
foreign key (page) references Page(id)
foreign key (member) references Member(id));
我正在使用spring JDbc模板。
My interface looks like follows: accountid is the memeberid in the database.
Collection<Diary> getDiaries(Long accountId);
我的日记如下:
public class Diary {
private Collection<Page> pages;
private Long id;
private LocalTime creationDate;
private String name;
private String description;
private ViewType type;
}
如果我想使用jdbc模板准备日记对象,我想知道查询现在的样子。也可以只触发一个查询并准备这个Diary对象,因为我会避免同一个请求多次触发查询。对于上面的接口,我很可能会使用s join命令,或者使用spring JDBC模板框架有更简单的方法。
答案 0 :(得分:3)
您可以使用外部联接创建单个查询(我假设这里可能有一个没有任何页面的日记,并且页面没有注释)。
现在,不是进行多个查询(每个页面一个),而是使用外部联接进行单个查询,以连接Diary
,Page
和Comment
。正如您在下面看到的,这将意味着日志和页面信息会多次返回,但我认为在进行多次数据库调用和一些冗余信息之间存在权衡。
class FullDiaryRowCallbackHandler implements RowCallbackHandler {
private Collection<Diary> diaries = new ArrayList<Diary>();
private Diary currentDiary = null;
private Page currentPage = null;
public void processRow(ResultSet rs) {
long diaryId = rs.getLong("d.id");
if (currentDiary == null || diaryId != currentDiary.getId()) {
currentDiary = new Diary();
currentPage = null;
diaries.add(currentDiary);
currentDiary.setId(diaryId);
currentDiary.setCreationDate(toLocalTime(rs.getTimestamp("d.creationDate")));
currentDiary.setDescription(rs.getString("d.description"));
...
}
long pageId = rs.getLong("p.id");
if (!rs.wasNull() && currentPage != null && currentPage.getId() != pageId) {
currentPage = new Page();
if (currentDiary.getPages() == null) {
currentDiary.setPages(new ArrayList<Page>());
}
currentDiary.getPages().add(currentPage);
currentPage.setId(pageId);
currentPage.setCreationDate(toLocalTime(rs.getTimestamp("p.creationDate")));
...
}
long commentId = rs.getLong("c.id");
if (!rs.wasNull() && currentPage != null) {
Comment comment = new Comment();
if (currentPage.getComments() == null) {
currentPage.setComments(new ArrayList<Comment>());
}
currentPage.getComments().add(comment);
comment.setId(commentId);
comment.setPostingDate(toLocalTime(rs.getTimestamp("c.postingDate")));
comment.setComment(rs.getString("c.comment"));
}
}
public Collection<Diary> getDiaries() {
return diaries;
}
}
FullDiaryRowCallbackHandler rowCallbackHandler = new FullDiaryRowCallbackHandler();
Collection<Diary> result = jdbcTemplate.query(
"select d.id, " +
"d.creationDate, " +
"d.description, " +
"p.id, " +
"p.creationDate, " +
"c.id, " +
"c.postingDate, " +
"c.comment " +
"from Diary d " +
"left outer join Page p on d.id = p.diary " +
"left outer join Comment c on p.id = c.page " +
"where d.member = ? " +
"order by d.id, p.id, c.id",
rowCallbackHandler,
myMemberId);
Collection<Diary> diariesForMember = rowCallbackHandler.getDiaries();
请注意,代码并不是特别漂亮,因为您必须处理结果集并在返回新页面时进行处理(这就是order by
子句很重要的原因),但这就是那种当他们急切地想要Hibernate之类的东西照顾你时(我不是说Hibernate更好或不好,因为我喜欢使用JdbcTemplate为我提供的查询提供控件,但是在填充对象图时,Hibernate(或JPA)会做很多繁重的工作。)
编辑:
修改后返回会员的所有日记,页面和评论。