Linux中定時執行DB2的存儲過程

一、需求說明

　　將DB2中的實時數據按天和小時匯總到匯總表中。

 

二、過程及步驟

　　第一步：編寫DB2存儲過程。

　　1. 在Dbvisualizer中新建SQL editor。

　　2. 代碼：

--/
CREATE PROCEDURE "NCPOLLDATA"."AGGREGATE_ENTITY_DAILY"
()
LANGUAGE SQL
SPECIFIC SQL160427123206300
BEGIN
  --declare varibles
  DECLARE V_PREFIX_TIME VARCHAR(7);
  DECLARE V_START_TIME VARCHAR(16);
  DECLARE V_END_TIME VARCHAR(16);
  DECLARE V_AGGREGATE_TIMESTAMP BIGINT DEFAULT 0;
  DECLARE V_YEAR_VALUE INTEGER DEFAULT 0;
  DECLARE V_MONTH_VALUE INTEGER DEFAULT 0;
  DECLARE V_AGGREGATE_TIME TIMESTAMP;
  --set varibles value
  SET V_PREFIX_TIME = ('1' || REPLACE(SUBSTR(CHAR(current_timestamp - 1 DAYS), 3, 9), '-', ''));
  SET V_START_TIME = V_PREFIX_TIME || '000000000';
  SET V_END_TIME = V_PREFIX_TIME || '235959999';
  SET V_YEAR_VALUE = YEAR(current_timestamp - 1 DAYS);
  SET V_MONTH_VALUE = MONTH(current_timestamp - 1 DAYS);
  SET V_AGGREGATE_TIME = current_timestamp - 1 DAYS;
  SET V_AGGREGATE_TIMESTAMP = TIMESTAMPDIFF(2, CHAR(current_timestamp - TIMESTAMP('1970-01-01-00.00.00')));
  --action
  FOR TEMP_CUR AS 
    SELECT t1.MONITOREDOBJECTID AS "POLICY_ID", t2.MAINNODEENTITYID AS "ENTITY_ID", 
    MIN(VALUE) AS "MIN_VALUE", MAX(VALUE) AS "MAX_VALUE",  AVG(BIGINT(VALUE)) AS "AVG_VALUE", 
    COUNT(VALUE) AS "DATA_COUNT", SUM(BIGINT(VALUE)) AS "DATA_SUM" 
    FROM POLLDATA t1 LEFT JOIN MONITOREDINSTANCE t2 ON t1.MONITOREDINSTID = t2.MONITOREDINSTID 
    WHERE t1.TDWTIME BETWEEN V_START_TIME AND V_END_TIME 
    GROUP BY t1.MONITOREDOBJECTID, t2.MAINNODEENTITYID 
    ORDER BY t1.MONITOREDOBJECTID, t2.MAINNODEENTITYID
  DO 
  --insert statement
    INSERT INTO POLLDATA_ENTITY_DAILY 
    VALUES (TEMP_CUR.POLICY_ID, TEMP_CUR.ENTITY_ID, V_AGGREGATE_TIMESTAMP, 
      V_YEAR_VALUE, V_MONTH_VALUE, V_AGGREGATE_TIME, 
      TEMP_CUR.MIN_VALUE, TEMP_CUR.MAX_VALUE, TEMP_CUR.AVG_VALUE, TEMP_CUR.DATA_COUNT, TEMP_CUR.DATA_SUM); 
  END FOR;
END
/

　　注意：在Dbvisualizer中編寫DB2存儲過程，第一行的 --/  和最后一行的 / 是必不可少的否則報錯。

　　3. 執行存儲過程: 點擊Dbvisualizer的存儲過程編輯框的執行按鈕即可，也可以在命令行中執行call procedure_name。

 

　　第二步：編寫可在Linux系統執行的腳本命令。

#!/bin/sh

DB2_HOME=/opt/IBM/tivoli/netcool/platform/linux2x86/db2

PATH=$DB2_HOME/bin:$PATH

export $DB2_HOME

export $PATH

startTime=`date +%Y%m%d%H%M%S`

echo $startTime

echo "開始執行腳本"

db2 connect to REPORTER user username using password

echo "連接上數據庫，開始執行腳本"

db2 "call NCPOLLDATA.AGGREGATE_ENTITY_DAILY()"

endTime=`date +%Y%m%d%H%M%S`

echo $endTime

echo "腳本執行完畢！關閉數據庫連接"

db2 terminate

 

　　以上命令參考資料：http://blog.csdn.net/sweetbowie/article/details/19614289

 

　　第三步：設置crond服務隨Linux系統自啟動

　　root@localhost ~]# /etc/init.d/crond restart

　　如果讓crond在開機時運行，應該改變其運行級別

　　[root@localhost ~]# chkconfig --levels 35 crond on

 

　　【相關資料鏈接】

　　1. http://blog.csdn.net/sweetbowie/article/details/19614289

　　2. http://www.linuxidc.com/Linux/2012-02/54391.htm

　　3. http://blog.csdn.net/liu_xing_hui/article/details/6565804/

　　4. http://blog.csdn.net/lulei9876/article/details/9170151

　　5. http://blog.csdn.net/jiujie395/article/details/8692655