hadoop hive 与 Oracle 互相导入数据

#1.在Oracle上建立要处理的表
create table SOURCE_TABLE_NAME as 
SELECT t.*,rownum as row_num FROM SOURCE_TABLE_NAME_O t ;
alter table SOURCE_TABLE_NAME
  add constraint SOURCE_TABLE_NAME_P primary key (ROW_NUM);
#2.在Oracle上建立处理结果表
DEST_TABLE_NAME
#3.oracle导入到hadoop
nohup 
sqoop import 
--hive-import 
--connect jdbc:oracle:thin:@192.168.1.14:1521:oradb1 
--username USER1 
--password "密码" 
--verbose -m 32 
--table SOURCE_TABLE_NAME 
--hive-table db_hive.SOURCE_TABLE_NAME 
--fields-terminated-by 't' 
--lines-terminated-by 'n' &
#4.计算
hive -e "drop table db_hive.DEST_TABLE_NAME ;
create table db_hive.DEST_TABLE_NAME row format delimited fields terminated by 't' STORED AS TEXTFILE as 
select 
max(Dn)  as  Dn, 
EutranCellTdd_uk , 
max(EutranCellTdd_name)  as  EutranCellTdd_name, 
max(GsmRelation) as GsmRelation, 
adj_uk, 
max(adj_name) as adj_name, 
max(EnbFunction_uk)  as  EnbFunction_uk, 
max(EnbFunction_name)  as  EnbFunction_name, 
max(ManagedElement_uk)  as  ManagedElement_uk, 
max(ManagedElement_name)  as  ManagedElement_name, 
max(omc_uk)  as  omc_uk, 
max(omc_name)  as  omc_name, 
sum(HO_ToGsmAttOutPerRelation)         as HO_ToGsmAttOutPerRelation, 
sum(HO_ToGsmSuccOutPrepPerRelation)    as HO_ToGsmSuccOutPrepPerRelation, 
sum(HO_ToGsmSuccOutPerRelation)        as HO_ToGsmSuccOutPerRelation, 
max(VENDOR_UK)  as  VENDOR_UK, 
max(VENDOR_NAME)  as  VENDOR_NAME, 
max(city_name) as city_name 
from db_hive.SOURCE_TABLE_NAME 
group by EutranCellTdd_uk, adj_uk ;"
#5.hadoop导出oracle（需要清空目标表DEST_TABLE_NAME）
sqoop export 
--connect jdbc:oracle:thin:@192.168.1.14:1521:oradb1 
--username USER1 
--password "密码" 
--table DEST_TABLE_NAME 
--export-dir /user/hive/warehouse/db_hive.db/DEST_TABLE_NAME 
--input-fields-terminated-by 't' 
--input-lines-terminated-by 'n' 
--null-string '\N' 
--null-non-string '\N'