CREATE EXTERNAL

TABLE IF NOT EXISTS data_zh(

ROWKEY STRING,

STATION INT,

YEAR INT,

MONTH INT,

DAY INT,

HOUR INT,

MINUTE INT,

)

PARTITIONED BY (AGE INT)

指定分区(此列并没真正存储列，也就是不存于你的数据中。但是如果你的数据从Oracle按年份导出，按照年份分区，把每一年数据放到age对应的目录下)

ROW FORMAT DELIMITED

FIELDS TERMINATED BY ‘,’

字段分隔符

LINES TERMINATED BY ‘\n’ 行分隔符

STORED AS TEXTFILE; 作为文本存储

Hive加载数据到表中

加载数据

LOAD DATA INPATH

‘/data/’ OVERWRITE INTO TABLE data_zh;

#加载某个目录下所有数据，存在分区数据不能这样加载

分区加载数据

创建分区

ALTER TABLE data_zhp ADD PARTITION(AGE= 1998)

加载数据

LOAD DATA INPATH ‘/data/1998.txt’ INTO TABLE data_zhp

PARTITION(YEAR=1998);

指定分区目录

LOCATION ‘

/hiveuser/hive/warehouse/data_zh.db/data_zh/2012.txt’;(指定分区所在位置)

加载数据到指定分区，分区表加载方法

下面为将现有表，修改为分区表。注意创建分区的时候未指定分区表，不能创建分区。需要新建表哥，用动态分区导入数据，动态分区数量有限。参考Programming

Hive记得最大上限应该是10000

Hive修改现有表添加分区

添加分区

ALTER TABLE data_zh ADD IF NOT EXISTS

PARTITION (year = 1998) LOCATION

‘/hiveuser/hive/warehouse/data_zh.db/data_zh/1998.txt’

删除分区

ALTER TABLE data_zhp DROP IF EXISTS PARTITION(year =1998);

修改现有表添加分区

创建分区表

CREATE EXTERNAL TABLE IF NOT EXISTS data_zhp(

ROWKEY STRING,

STATION INT,

MONTH INT,

DAY INT,

HOUR INT,

MINUTE INT,

)

PARTITIONED BY (YEAR INT)

ROW FORMAT DELIMITED

FIELDS TERMINATED BY ‘,’

LINES TERMINATED BY ‘\n’

STORED AS TEXTFILE;

动态分区指令

set

hive.exec.dynamic.partition=true;

set hive.exec.dynamic.partition.mode=nonstrict;

#set hive.enforce.bucketing = true;

未分区表数据导入分区表

insert overwrite table data_zhp partition (YEAR,MONTH) select *

from data_zh;