OLAP的第一步就是从业务系统中抽取数据到数据仓库系统。
除了ETL工具Kettle，也可以使用PL/SQL

创建Source表，模拟业务系统的数据表。
create table source as select sys_guid() id ,o.* from dba_objects o where rownum<1000;
alter table source add constraint PK_source primary key (id);

创建Target表，模拟数据仓库中的表。
create table target as select * from source where 1=0;
alter table target add constraint PK_target primary key (id);

创建DB LINK，remote模拟业务系统的数据库
create database link remote connect to username identified by xxxxxx using 'remote';

因为业务系统的数据是变化的，相较于数据仓库的表，他可能更新了一些数据，也可能修改了一些数据。
比对业务表和数据仓库表的数据，
如果ID相同，并且数据有变化
      则根据ID更新数据仓库的表(target)
如果业务系统的数据ID(source)，还没有出现在数据仓库的表中(target)
      则在数据仓库的表中新增这个记录。

本质都是Oracle Merge的功能，只不过尝试另外几种方法。

1.merge

merge into target t using (select * from source@remote) s

on(t.id=s.id)

when matched

then

    update set t.owner=s.owner,

                t.object_name=s.object_name,

                t.subobject_name=s.subobject_name,

                t.object_id=s.object_id,

                t.data_object_id=s.data_object_id,

                t.object_type=s.object_type,

                t.created=s.created,

                t.last_ddl_time=s.last_ddl_time,

                t.timestamp=s.timestamp,

                t.status=s.status,

                t.temporary=s.temporary,

                t.generated=s.generated,

                t.secondary=s.secondary,

                t.namespace=s.namespace,

                t.edition_name=s.edition_name

when not matched

then

    insert values

    (

                s.id,

                s.owner,

                s.object_name,

                s.subobject_name,

                s.object_id,

                s.data_object_id,

                s.object_type,

                s.created,

                s.last_ddl_time,

                s.timestamp,

                s.status,

                s.temporary,

                s.generated,

                s.secondary,

                s.namespace,

                s.edition_name

    );

2.全局临时表。
    首先将远程业务系统的数据放入临时表，
然后根据ID更新数据，如果数据的内容没有变化，则不更新。
最后插入业务系统中新建的数据。

create global temporary table tmp

on commit preserve rows

as

select * from target where 1=0;

insert into tmp select * from source@remote;

update target t set

            (

                t.owner,

                t.object_name,

                t.subobject_name,

                t.object_id,

                t.data_object_id,

                t.object_type,

                t.created,

                t.last_ddl_time,

                t.timestamp,

                t.status,

                t.temporary,

                t.generated,

                t.secondary,

                t.namespace,

                t.edition_name

            )

            =

            (select

                tmp.owner,

                tmp.object_name,

                tmp.subobject_name,

                tmp.object_id,

                tmp.data_object_id,

                tmp.object_type,

                tmp.created,

                tmp.last_ddl_time,

                tmp.timestamp,

                tmp.status,

                tmp.temporary,

                tmp.generated,

                tmp.secondary,

                tmp.namespace,

                tmp.edition_name

            from tmp where t.id=tmp.id)

    where exists(

                    select * from tmp where tmp.id=t.id and not (

                                tmp.owner=t.owner and

                                tmp.object_name=t.object_name and

                                tmp.subobject_name=t.subobject_name and

                                tmp.object_id=t.object_id and

                                tmp.data_object_id=t.data_object_id and

                                tmp.object_type=t.object_type and

                                tmp.created=t.created and

                                tmp.last_ddl_time=t.last_ddl_time and

                                tmp.timestamp=t.timestamp and

                                tmp.status=t.status and

                                tmp.temporary=t.temporary and

                                tmp.generated=t.generated and

                                tmp.secondary=t.secondary and

                                tmp.namespace=t.namespace and

                                tmp.edition_name=t.edition_name

                    )

                );

insert into target

    select * from tmp where not exists(

        select * from target t where t.id=tmp.id);

3.集合处理
为了简单，没有进行内容变化的判断

declare

    type tab is table of target%rowtype;

    l_row tab;

    cursor cur is select * from source@remote;

begin

    open cur;

    fetch cur bulk collect into l_row;

    close cur;

    forall i in 1..l_row.count

        update target set row=l_row(i) where id=l_row(i).id;

    insert into target select * from source@remote s

        where not exists (select * from target t where t.id=s.id);

    commit;

end;

/

4.内联视图更新
没有写更新后插入的步骤，插入的实现是相同的。

update(

    select

        s.id s1,

        s.owner s2,

        s.object_name s3,

        s.subobject_name s4,

        s.object_id s5,

        s.data_object_id s6,

        s.object_type s7,

        s.created s8,

        s.last_ddl_time s9,

        s.timestamp s10,

        s.status s11,

        s.temporary s12,

        s.generated s13,

        s.secondary s14,

        s.namespace s15,

        s.edition_name s16,

        t.id t1,

        t.owner t2,

        t.object_name t3,

        t.subobject_name t4,

        t.object_id t5,

        t.data_object_id t6,

        t.object_type t7,

        t.created t8,

        t.last_ddl_time t9,

        t.timestamp t10,

        t.status t11,

        t.temporary t12,

        t.generated t13,

        t.secondary t14,

        t.namespace t15,

        t.edition_name t16

    from target t inner join source@remote s on(s.id=t.id)

    where

        not

        (

            s.owner=t.owner and

            s.object_name=t.object_name and

            s.subobject_name=t.subobject_name and

            s.object_id=t.object_id and

            s.data_object_id=t.data_object_id and

            s.object_type=t.object_type and

            s.created=t.created and

            s.last_ddl_time=t.last_ddl_time and

            s.timestamp=t.timestamp and

            s.status=t.status and

            s.temporary=t.temporary and

            s.generated=t.generated and

            s.secondary=t.secondary and

            s.namespace=t.namespace and

            s.edition_name=t.edition_name

        )

)

set

    t1=s1,

    t2=s2,

    t3=s3,

    t4=s4,

    t5=s5,

    t6=s6,

    t7=s7,

    t8=s8,

    t9=s9,

    t10=s10,

    t11=s11,

    t12=s12,

    t13=s13,

    t14=s14,

    t15=s15,

    t16=s16

;

5.Minus
先插入业务表中新增的记录，然后对比修改。

declare

    type tab is table of target%rowtype;

    l_row tab;

    cursor cur is select * from source@remote minus select * from target;

begin

    insert into target select * from source@remote s

        where not exists (select * from target t where t.id=s.id);

    open cur;

    fetch cur bulk cllect into l_row;

    close cur;

    forall i in 1..l_row.count

        update target set row=low(i) where id=l_row(i).id;

    commit;

end;

/

ITPub博客

在两个数据库之间进行数据同步

最新文章