SQL语句性能调整一例

公司开发同事发来一条语句，反应执行很慢，原语句如下：

SELECT COUNT(df.origrec)* it.scale AS pcount,
it.invcode                       AS invcode,
df.samplefrom                    AS rasclientid
FROM da_finance_pricesett df,
dbs_db_invbtest it
WHERE df.testcode                      = it.testcode
AND TO_CHAR(df.senddate,'yyyy-mm-dd')  > '2013-02-25'
AND TO_CHAR(df.senddate,'yyyy-mm-dd') <= '2013-02-25'
AND it.invcode                        IN
(SELECT DISTINCT invcode FROM dbs_db_invbtest
)
GROUP BY it.scale,
it.invcode,
df.samplefrom;

存在的问题：

1、条件子句条件含函数引起的隐性转换导致全表扫描；

2、条件子句的嵌套条件存在无条件过滤导致全表扫描,而且从连接条件逻辑看，该子句与第一次等值连接是重复的。修改如下：

SELECT COUNT(df.origrec)* it.scale AS pcount,
it.invcode AS invcode,
df.samplefrom  AS rasclientid
FROM   da_finance_pricesett df,
dbs_db_invbtest it
WHERE  df.testcode = it.testcode
AND  df.senddate   >to_date('2013-01-25','yyyy-mm-dd')
AND  df.senddate  <=to_date('2013-02-25','yyyy-mm-dd')
GROUP BY it.scale,
it.invcode,
df.samplefrom;