1、建表，
=========================================================
create table T_JD_RECOMM(
TIP_NO varchar2(6),
USR_NO VARCHAR2(10),
SKU_NO VARCHAR2(15),
TRDTYPE_NO VARCHAR2(6),
TRDTYPE_NAME VARCHAR2(20)
);
==========================================================
2、写ctl:
===========================================================
load data
infile 'G:PROGRAM .自主开发jdjd_test.csv'
into table T_JD_RECOMM
fields terminated by ','
(TIP_NO,USR_NO,SKU_NO,TRDTYPE_NO,TRDTYPE_NAME)
============================================================
3、sqlldr userid=system/jianxiao@orcl control=G:PROGRAM .自主开发jdld.ctl log=G:PROGRAM .自主开发jdld.log
其它方法：
1. A 表的记录导出为一条条分号隔开的 insert 语句，然后执行插入到 B 表中
2. 建立数据库间的 dblink，然后用 create table B as select * from A@dblink where ...，或 insert into B select * from A@dblink where ...
3. exp A 表，再 imp 到 B 表，exp 时可加查询条件
4. 程序实现 select from A ..，然后 insert into B ...，也要分批提交
5. 再就是本篇要说到的 Sql Loader(sqlldr) 来导入数据，效果比起逐条 insert 来很明显