LIMS系统ODI配置与使用

Posted 2020-08-09

以食品REPORTS表为例，说明ODI在LIMS系统中的配置与使用。

先为REPORTS表添加列LIMS_CODE

注意：类型为NUMBER(*,0)，不允许为空，默认值为4（中心为1、可靠性为2、环境为3、食品为4）；

说明：

    ODI的增量刷新要求源表与目标表都有主键，并且同步根据主键来刷新其他字段，这就要求主键值永久不变。在LIMS系统中，主键值是会变的，比如ORDTASK中的，就有修改测试方法的功能，如果主键发生了变化，则后续的刷新就无法完成。为了使LIMS能够正常同步，引入了ORIGRE + LIMS_CODE组合，以确保在系统中唯一，同时也保证值永久不会变化。

再添加唯一索引

在模型ODI_SRC_FDD（为每套系统建立起一套模型）中，添加REPORTS的元数据信息(配置模型等过程省略)

如下图，打开ODI_SRC_FDD模型

选择右边的取消全选

选择REPORTS，再选择“Reverse Engineer”

选择“Yes”

逆向工程完成后，ODI_SRC_FDD模型中就有了ORDTASK的信息，如图

接下来，需要配置同步信息了。为了简化配置操作，封装了函数CTI_FDD_DATA.FN_ODI，并且创建了公共同义词

如上图，第一个参数为模式名，USER为当前模式名，REPORTS为表名，函数返回信息如下：

INSERT INTO SNP_SUBSCRIBERS(JRN_TNAME,JRN_SUBSCRIBER,JRN_REFDATE,JRN_ROW_COUNT,JRN_DATA_CMD,JRN_COUNT_CMD)
               VALUES (‘CTI_FDD_DATA.REPORTS‘,‘REPORTS‘,SYSDATE,0,NULL,‘update CTI_FDD_DATA.SNP_SUBSCRIBERS    SUBS set SUBS.JRN_ROW_COUNT=(select count(‘‘X‘‘) from CTI_FDD_DATA.J$REPORTS    JRN where JRN.JRN_SUBSCRIBER=‘‘$$SUBSCRIBER_NAME$$‘‘) where SUBS.JRN_SUBSCRIBER=‘‘$$SUBSCRIBER_NAME$$‘‘ and SUBS.JRN_TNAME = ‘‘CTI_FDD_DATA.REPORTS‘‘‘);
commit;

CREATE TABLE J$REPORTS(jrn_subscriber VARCHAR2(400 CHAR) not null,
                                                 jrn_consumed   VARCHAR2(1 CHAR),
                                                 jrn_flag       VARCHAR2(1 CHAR),
                                                 jrn_date       DATE,
                                                 ORIGREC        NUMBER(*,0),
                                                 LIMS_CODE      NUMBER(*,0));

create or replace trigger T$REPORTS
  after insert or update or delete on CTI_FDD_DATA.REPORTS
  for each row
declare
  V_FLAG     VARCHAR(1);
  V_ORIGREC  NUMBER;
  V_LIMS_CODE NUMBER;
begin
  if inserting then
    V_ORIGREC    := :new.ORIGREC;
    V_LIMS_CODE  := :new.LIMS_CODE;
    V_FLAG     := ‘I‘;
  end if;

  if updating then
    V_ORIGREC    := :new.ORIGREC;
    V_LIMS_CODE  := :new.LIMS_CODE;
    V_FLAG     := ‘I‘;
  end if;

  if deleting then
    V_ORIGREC    := :old.ORIGREC;
    V_LIMS_CODE  := :old.LIMS_CODE;
    V_FLAG     := ‘D‘;
  end if;

  insert into CTI_FDD_DATA.J$REPORTS
    (JRN_SUBSCRIBER,
     JRN_CONSUMED,
     JRN_FLAG,
     JRN_DATE,
     ORIGREC,
     LIMS_CODE)
    select JRN_SUBSCRIBER,
           ‘0‘,
           V_FLAG,
           sysdate,
           V_ORIGREC,
           V_LIMS_CODE
      from CTI_FDD_DATA.SNP_SUBSCRIBERS
     where JRN_TNAME = ‘CTI_FDD_DATA.REPORTS‘;
end;

create or replace view jv$dREPORTS as
    select decode(TARG.ROWID, null, ‘D‘, ‘I‘) AS JRN_FLAG,
       JRN.JRN_SUBSCRIBER AS JRN_SUBSCRIBER,
       JRN.JRN_DATE AS JRN_DATE,
       JRN.ORIGREC AS ORIGREC,
       JRN.LIMS_CODE AS  LIMS_CODE, TARG.ORIGSTS,TARG.FOLDERNO,TARG.REPORTNO,TARG.REPORTTYPE,TARG.REPORTCAT,TARG.REPORTGENTYPE,TARG.STATUS,TARG.DISPSTS,TARG.ASSESS,TARG.ASSESS_ENG,TARG.REMARK,TARG.ADDEDBY,TARG.ADDEDDATE,TARG.TESTBY,TARG.TESTBYFULLNAME,TARG.TESTDATE,TARG.APPROVEDBY,TARG.APPROVEDDATE,TARG.RELEASEDBY,TARG.RELEASEDDATE,TARG.PATH,TARG.PATH_PDF,TARG.PATH_WATERMARK,TARG.DIVISIONCODE,TARG.DEPT,TARG.DEPTNAME,TARG.ORDNOS,TARG.REFUSEFLAG,TARG.EDIT_TIMES,TARG.SENDEXPRESSNO,TARG.NEEDSIGN,TARG.HASSIGNED,TARG.ALLOWEDIT,TARG.HASSIGNED_W
       from (select L.JRN_SUBSCRIBER AS JRN_SUBSCRIBER,
               L.ORIGREC AS ORIGREC,
               L.LIMS_CODE AS LIMS_CODE,
               max(L.JRN_DATE) AS JRN_DATE
          from CTI_FDD_DATA.J$REPORTS L
         group by L.JRN_SUBSCRIBER, L.ORIGREC,L.LIMS_CODE) JRN,
       CTI_FDD_DATA.REPORTS TARG
 where JRN.ORIGREC = TARG.ORIGREC(+)
   and JRN.LIMS_CODE = TARG.LIMS_CODE(+);

create or replace view jv$REPORTS as
select   decode(TARG.ROWID, null, ‘D‘, ‘I‘)   AS  JRN_FLAG,
  JRN.JRN_SUBSCRIBER     AS  JRN_SUBSCRIBER,
  JRN.JRN_DATE     AS  JRN_DATE,
  JRN.ORIGREC AS ORIGREC,
  JRN.LIMS_CODE AS  LIMS_CODE,TARG.ORIGSTS,TARG.FOLDERNO,TARG.REPORTNO,TARG.REPORTTYPE,TARG.REPORTCAT,TARG.REPORTGENTYPE,TARG.STATUS,TARG.DISPSTS,TARG.ASSESS,TARG.ASSESS_ENG,TARG.REMARK,TARG.ADDEDBY,TARG.ADDEDDATE,TARG.TESTBY,TARG.TESTBYFULLNAME,TARG.TESTDATE,TARG.APPROVEDBY,TARG.APPROVEDDATE,TARG.RELEASEDBY,TARG.RELEASEDDATE,TARG.PATH,TARG.PATH_PDF,TARG.PATH_WATERMARK,TARG.DIVISIONCODE,TARG.DEPT,TARG.DEPTNAME,TARG.ORDNOS,TARG.REFUSEFLAG,TARG.EDIT_TIMES,TARG.SENDEXPRESSNO,TARG.NEEDSIGN,TARG.HASSIGNED,TARG.ALLOWEDIT,TARG.HASSIGNED_W
  from  (
    select  L.JRN_SUBSCRIBER   AS  JRN_SUBSCRIBER,
      L.ORIGREC AS ORIGREC,
      L.LIMS_CODE AS LIMS_CODE,
      max(L.JRN_DATE)  AS  JRN_DATE
    from  CTI_FDD_DATA.J$REPORTS    L
    where L.JRN_CONSUMED =‘1‘
    group by  L.JRN_SUBSCRIBER,
        L.ORIGREC,L.LIMS_CODE
  )    JRN,
  CTI_FDD_DATA.REPORTS    TARG
where JRN.ORIGREC = TARG.ORIGREC(+)
   and JRN.LIMS_CODE = TARG.LIMS_CODE(+);

执行完上面语句后，同步的基本信息已经配置完成。接下来，需要将表添加到CDC，再配置映射关系。

如图，选择“Add to CDC”

发现，为表REPORT添加了绿色小时钟。

新建映射

如图，将ODI_SRC_FDD下面的REPORTS表、ODI_TAG_DC下面的REPORTS表拖动到窗口中

从源表拉动到目标表，建立字段映射关系，如下图

点击工具栏上的勾，进行映射验证（主要验证类型及字段长度）

验证成功

修改订阅用户（由于我们的基础信息由自定义函数生成，订阅用户默认与表名相同）

选择“增量更新”

选择仅记录的日志

选择加载知识模型

选择集成知识模型

REPORTS_FDD映射已经配置完成

选择“Run”，验证执行情况

从执行日志上面看，配置是没有问题的。接下来，进行数据方面的测试

如图，向源表插入了一条记录，然后执行同步

查询目标表，数据已经刷新

修改数据（主键REPORTNO + 其他字段REPORTTYPE）并提交，执行同步

数据已经刷新

删除数据提交，并执行同步

数据已刷新。

 

附FN_ODI函数

CREATE OR REPLACE FUNCTION FN_ODI(v_SCHEMA VARCHAR2, v_TABLENAME VARCHAR2)
  RETURN CLOB AUTHID CURRENT_USER IS
  m_out  CLOB;
  m_cols CLOB;
  m_cnt  INTEGER;
BEGIN
  m_out := ‘‘;
  --创建订阅用户
  SELECT COUNT(*)
    INTO m_cnt
    FROM SNP_SUBSCRIBERS
   WHERE JRN_TNAME = v_SCHEMA || ‘.‘ || v_TABLENAME;
  IF m_cnt = 0 THEN
    m_out := ‘INSERT INTO SNP_SUBSCRIBERS(JRN_TNAME,JRN_SUBSCRIBER,JRN_REFDATE,JRN_ROW_COUNT,JRN_DATA_CMD,JRN_COUNT_CMD)
               VALUES (‘‘‘ || v_SCHEMA || ‘.‘ || v_TABLENAME ||
             ‘‘‘,‘‘‘ || v_TABLENAME || ‘‘‘,SYSDATE,0,NULL,‘‘update ‘ ||
             v_SCHEMA ||
             ‘.SNP_SUBSCRIBERS    SUBS set SUBS.JRN_ROW_COUNT=(select count(‘‘‘‘X‘‘‘‘) from ‘ ||
             v_SCHEMA || ‘.J$‘ || v_TABLENAME ||
             ‘    JRN where JRN.JRN_SUBSCRIBER=‘‘‘‘$$SUBSCRIBER_NAME$$‘‘‘‘) where SUBS.JRN_SUBSCRIBER=‘‘‘‘$$SUBSCRIBER_NAME$$‘‘‘‘ and SUBS.JRN_TNAME = ‘‘‘‘‘ ||
             v_SCHEMA || ‘.‘ || v_TABLENAME || ‘‘‘‘‘‘‘);‘;
    m_out := m_out || chr(10);
    m_out := m_out || ‘commit;‘;
    m_out := m_out || chr(10) || chr(10);
  END IF;

  -- 创建日志表
  m_out := m_out || ‘CREATE TABLE J$‘ || v_TABLENAME ||
           ‘(jrn_subscriber VARCHAR2(400 CHAR) not null,
                                                 jrn_consumed   VARCHAR2(1 CHAR),
                                                 jrn_flag       VARCHAR2(1 CHAR),
                                                 jrn_date       DATE,
                                                 ORIGREC        NUMBER(*,0),
                                                 LIMS_CODE      NUMBER(*,0));‘;

  --创建触发器
  m_out := m_out || chr(10) || chr(10);
  m_out := m_out || ‘create or replace trigger T$‘ || v_TABLENAME || ‘
  after insert or update or delete on ‘ || v_SCHEMA || ‘.‘ ||
           v_TABLENAME || ‘
  for each row
declare
  V_FLAG     VARCHAR(1);
  V_ORIGREC  NUMBER;
  V_LIMS_CODE NUMBER;
begin
  if inserting then
    V_ORIGREC    := :new.ORIGREC;
    V_LIMS_CODE  := :new.LIMS_CODE;
    V_FLAG     := ‘‘I‘‘;
  end if;

  if updating then
    V_ORIGREC    := :new.ORIGREC;
    V_LIMS_CODE  := :new.LIMS_CODE;
    V_FLAG     := ‘‘I‘‘;
  end if;

  if deleting then
    V_ORIGREC    := :old.ORIGREC;
    V_LIMS_CODE  := :old.LIMS_CODE;
    V_FLAG     := ‘‘D‘‘;
  end if;

  insert into ‘ || v_SCHEMA || ‘.J$‘ || v_TABLENAME || ‘
    (JRN_SUBSCRIBER,
     JRN_CONSUMED,
     JRN_FLAG,
     JRN_DATE,
     ORIGREC,
     LIMS_CODE)
    select JRN_SUBSCRIBER,
           ‘‘0‘‘,
           V_FLAG,
           sysdate,
           V_ORIGREC,
           V_LIMS_CODE
      from ‘ || v_SCHEMA || ‘.SNP_SUBSCRIBERS
     where JRN_TNAME = ‘‘‘ || v_SCHEMA || ‘.‘ || v_TABLENAME || ‘‘‘;
end;‘;

  --获取列信息
  SELECT WM_CONCAT(‘TARG.‘ || COLUMN_NAME)
    INTO m_cols
    FROM (SELECT COLUMN_NAME
            FROM USER_TAB_COLUMNS
           WHERE TABLE_NAME = v_TABLENAME
             AND COLUMN_NAME NOT IN (‘ORIGREC‘, ‘LIMS_CODE‘)
           ORDER BY COLUMN_ID);

  --创建视图
  m_out := m_out || chr(10) || chr(10);
  m_out := m_out || ‘create or replace view jv$d‘ || v_TABLENAME || ‘ as
    select decode(TARG.ROWID, null, ‘‘D‘‘, ‘‘I‘‘) AS JRN_FLAG,
       JRN.JRN_SUBSCRIBER AS JRN_SUBSCRIBER,
       JRN.JRN_DATE AS JRN_DATE,
       JRN.ORIGREC AS ORIGREC,
       JRN.LIMS_CODE AS  LIMS_CODE, ‘ || m_cols || ‘
       from (select L.JRN_SUBSCRIBER AS JRN_SUBSCRIBER,
               L.ORIGREC AS ORIGREC,
               L.LIMS_CODE AS LIMS_CODE,
               max(L.JRN_DATE) AS JRN_DATE
          from ‘ || v_SCHEMA || ‘.J$‘ || v_TABLENAME || ‘ L
         group by L.JRN_SUBSCRIBER, L.ORIGREC,L.LIMS_CODE) JRN,
       ‘ || v_SCHEMA || ‘.‘ || v_TABLENAME ||
           ‘ TARG
 where JRN.ORIGREC = TARG.ORIGREC(+)
   and JRN.LIMS_CODE = TARG.LIMS_CODE(+);‘;

  m_out := m_out || chr(10) || chr(10);
  m_out := m_out || ‘create or replace view jv$‘ || v_TABLENAME || ‘ as
select     decode(TARG.ROWID, null, ‘‘D‘‘, ‘‘I‘‘)     AS  JRN_FLAG,
    JRN.JRN_SUBSCRIBER         AS  JRN_SUBSCRIBER,
    JRN.JRN_DATE         AS  JRN_DATE,
  JRN.ORIGREC AS ORIGREC,
  JRN.LIMS_CODE AS  LIMS_CODE,‘ || m_cols || ‘
  from  (
    select  L.JRN_SUBSCRIBER   AS  JRN_SUBSCRIBER,
      L.ORIGREC AS ORIGREC,
      L.LIMS_CODE AS LIMS_CODE,
      max(L.JRN_DATE)  AS  JRN_DATE
    from  ‘ || v_SCHEMA || ‘.J$‘ || v_TABLENAME || ‘    L
    where L.JRN_CONSUMED =‘‘1‘‘
    group by  L.JRN_SUBSCRIBER,
        L.ORIGREC,L.LIMS_CODE
  )    JRN,
  ‘ || v_SCHEMA || ‘.‘ || v_TABLENAME ||
           ‘    TARG
where JRN.ORIGREC = TARG.ORIGREC(+)
   and JRN.LIMS_CODE = TARG.LIMS_CODE(+);‘;

  RETURN m_out;

END;