Hive使用HDFS目录数据创建Hive表分区

描述：

Hive表pms.cross_sale_path建立以日期作为分区，将hdfs目录/user/pms/workspace/ouyangyewei/testUsertrack/job1Output/crossSale上的数据，写入该表的$yesterday分区上

表结构：

hive -e " 
set mapred.job.queue.name=pms; 
 
drop table if exists pms.cross_sale_path; 
create external table pms.cross_sale_path 
( 
track_id string, 
track_time string, 
session_id string, 
gu_id string, 
end_user_id string, 
page_category_id bigint, 
algorithm_id int, 
is_add_cart int, 
rcmd_product_id bigint, 
product_id bigint, 
category_id bigint, 
path_id string, 
path_type int, 
path_length int, 
path_list string, 
order_code string 
) 
PARTITIONED BY (ds string) 
ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t' 
LINES TERMINATED BY '\n' 
STORED AS TEXTFILE;" 

写法：

yesterday=`date -d -1days +%Y-%m-%d` 
hive -e "load data inpath '/user/pms/workspace/ouyangyewei/testUsertrack/job1Output/crossSale' into table pms.cross_sale_path partition(ds='$yesterday');" 

Hive 的详细介绍：请点这里
Hive 的下载地址：请点这里