�ݺ�ߣ

ORM 을 맞이하는  
우리의 자세
정병태

랑겔한스 : 정병태
- Web Developer 
- OKKY.KR Contributor 
- https://github.com/1angerhans
- http://okky.kr/user/info/26138
- btjung@ebrain.kr

‘나는 SQL 이 싫어요.’

ORM 
Object Relational Mapping

– Wikipedia 인용
“Object-relational mapping (ORM, O/RM, and O/R
mapping) in computer science is a programming
technique for converting data between
incompatible type systems in object-oriented
programming languages.”

– Wikipedia 인용
“객체 관계 매핑(Object-relational mapping; ORM)은
데이터베이스와 객체 지향 프로그래밍 언어 간의 호환되
지 않는 데이터를 변환하는 프로그래밍 기법이다.”

Object-relational
impedance mismatch
- Data type differences
- Structural and integrity differences
- Manipulative difference
- Transactional differences
https://en.wikipedia.org/wiki/Object-relational_impedance_mismatch

Vo, DTO 그리고 �ղ��들

CREATE TABLE TBL_MSG
(
F_GRP_ID INT(6) NOT NULL PRIMARY KEY ,
PTN_ID INT(6) ,
PRIORITY INT(6) ,
JOB_CODE INT(6) ,
FLAG_PRO INT(6) ,
CHAINNUM INT(6) ,
MASSNUM INT(6) ,
USERID VARCHAR(20) ,
CPCODE VARCHAR(20) ,
QTCODE VARCHAR(20) ,
MSGSEPER CHAR(1) ,
MSG_CNT INT(6) ,
흔한 Legacy Table

CREATE TABLE TBL_USER (
SEQ INT(6) NOT NULL PRIMARY KEY ,
USER_TYPE VARCHAR(20) ,
AUTH_TYPE VARCHAR(20) ,
NAME VARCHAR(20) ,
PASS VARCHAR(20) ,
TEL VARCHAR(12) ,
CELL VARCHAR(12) ,
OFFICE_TEL VARCHAR(12) ,
EMAIL VARCHAR(200) ,
ZIPCODE VARCHAR(7) ,
ADDRESS_01 VARCHAR(200) ,
COMPANY_NAME VARCHAR(20) ,
QUARTER_NAME VARCHAR(20) ,
STATUS VARCHAR(20) ,
MILEAGE INT(12) ,
REG_DATE DATE ,
PASS_CHG_DATE DATE ,
PASS_FAIL_CNT INT(6)
);
User
Contact
Company
Domains
Auth
Mileage

CREATE TABLE TBL_USER (
USER_TYPE VARCHAR(20) ,
NAME VARCHAR(20) ,
PASS VARCHAR(20) ,
REG_DATE DATE ,
PASS_CHG_DATE DATE ,
PASS_FAIL_CNT INT(6)
);
CREATE TABLE TBL_COMPANY (
COMPANY_NAME VARCHAR(20) ,
REG_DATE DATE
);
CREATE TABLE TBL_AUTH (
ROLE VARCHAR(20)
);
CREATE TABLE TBL_ADDRESS (
ZIPCODE VARCHAR(7) ,
ADDRESS_02 VARCHAR(200)
);
CREATE TABLE TBL_CONTACT (
TEL VARCHAR(12) ,
CELL VARCHAR(12)
);
CREATE TABLE TBL_MILEAGE (
POINT INT(12)
);

public class User {
Long seq;
String userid;
String userpass;
String username;
Date regDate;
Company company;
Address address;
Contact contact;
Mileage mileage;
List<Auth> authorities;
// Getter & Setter
}
public class Company {
…
}
public class Address {
…
}
public class Contact {
…
}
public class Mileage {
…
}
public class Auth {
…
}

DBA 들이 나를 향해 이렇게 외친다.

public class MSG {
int grpnum;
int chainnum;
String userid;
String usernm;
String cpcode;
String qtcode;
String title;
String sndTitle;
String cComment;
String dstaddr;
String callback;
String msgTemplet;
흔한 Legacy Model

<select id="selectOne" resultMap=“userMap">
select
u.seq, u.userid, u.pass, u.name, u.tel, u.cel, u.email, u.company_name, u.zipcode,
u.address_01, u.address_02 ….
from
tbl_user u
where
u.seq = #{seq}
</select>
Mybatis

Mybatis Association
<resultMap id="userMap" type=“User">
<id jdbcType="INTEGER" property="seq" column="seq"/>
<result jdbcType="VARCHAR" property="userid" column="userid"/>
<result jdbcType="VARCHAR" property="pass" column="pass"/>
<association property="company" javaType="Company" >
<result jdbcType="VARCHAR" property=“company_name" column="cpname"/>
</association>
<association property="contact" javaType="Contact" >
<result jdbcType="VARCHAR" property=“company_name" column="cpname"/>
</association>
<association property="address" javaType="Address" >
<result jdbcType="VARCHAR" property=“zopcode" column=“zopcode"/>
<result jdbcType="VARCHAR" property=“address_01” column=“address_01"/>
<result jdbcType="VARCHAR" property=“address_01” column="address_01"/>
</association>
<collection property="authorities" select="selectUserAuthList" fetchType="lazy" column="{seq=seq}" />
</resultMap>

– 미상
“인간의 욕심은 끝이 없고, 같은 실수를 반복한다.”

@Entity
@Table( name = “TBL_USER" )
public class User {
Integer seq;
String userid;
String userpass;
String username;
Date regDate;
Company company;
Address address;
Contact contact;
Mileage mileage;
// Getter & Setter
@Id
@GeneratedValue(generator="increment")
@GenericGenerator(name="increment", strategy = “increment”)
@Column(name = “SEQ”)
Long getSeq() {
return seq;
}
}
@Entity
@Table( name = “TBL_COMPANY" )
…
}
@Entity
@Table( name = “TBL_ADDRESS" )
…
}
@Entity
@Table( name = “TBL_CONTACT" )
…
}
@Entity
@Table( name = “TBL_MILEAGE" )
…
}
@Entity
@Table( name = “TBL_AUTH" )
public class Auth {
…
}

@Entity
@Table( name = “TBL_USER" )
public class User {
@Id
@GeneratedValue(generator="increment")
@GenericGenerator(name="increment", strategy = “increment”)
@Column(name = “SEQ”)
Integer seq;
@Column(name = “USERID”)
String userid;
@Embedded
Company company;
@Embedded
Address address;
@Embedded
Contact contact;
@OneToOne( cascade = CascadeType.ALL, fetch = FetchType.LAZY)
Mileage mileage;
@OneToMany( cascade = CascadeType.ALL, fetch = FetchType.LAZY)
}
@Embeddable
…
}
@Embeddable
…
}
@Embeddable
…
}
@Embeddable
…
}
@Embeddable
public class Auth {
…
}

Criteria
Criteria criteria = session.createCriteria(User.class);
criteria.add(Restrictions.eq("seq", seq))
.setProjection(
Projections.projectionList()
.add(Projections.property(“zipcode"))
.add(Projections.property(“address_01"))
.add(Projections.property(“address_02")))
.setResultTransformer(
Transformers.aliasToBean(Address.class));
Address address = criteria.uniqueResult();

HQL
Address address = session.createQuery(“select u.zipcode,
u.address_01, u.address_02 from User u where seq = :seq”, )
.setParameter("seq", seq)
.setResultTransformer(
Transformers.aliasToBean(Address.class))
.uniqueResult();

ORM 에 대한 단상
• Legacy DB 도 충분히 도입 가능하다.
• ORM을 염두한 디자인을 적용할 경우 생산성이 급격히 좋아진다.
• Criteria는 학습이 쉽지 않다.
• HQL을 쓸 때면 뭔가 손해보는 기분.. (방언 처리 라고는 하지만..)
• 퍼포먼스는 사실상 문제 없다.
• SQL Layer가 빠져야돼.

ORM 은
우리가 가지고 있는 문제를 해결한
‘결과’가 아닌
더욱 나은 방향으로 가기 위한
‘과정’ 입니다.
– 저

SQL이 필요 없다고
RDBMS가 필요 없는건 아닐껍니다.

�ݺ�ߣ

ORM을 맞이하는 우리의 자세

Recommended

More Related Content

Similar to ORM을 맞이하는 우리의 자세 (20)

Recently uploaded (20)

ORM을 맞이하는 우리의 자세