HIVE SQL 實操

基本操作

建表

show databases;

show tables;

create database mydb;

create table if not exists 
mydb.employee(
eid int, 
name String, 
salary String, 
destination String) 
COMMENT 'Employee details' 
ROW FORMAT DELIMITED 
FIELDS TERMINATED BY '\t' 
LINES TERMINATED BY '\n' STORED AS TEXTFILE;

描述表，查看schema

DESCRIBE mydb.employee;

從文件導入數據

load data local inpath '/root/text.csv' into table employee;

hive 中執行 shell 命令

!clear; //hive中執行shell命令, 清空屏幕

hive 中執行 hdfs 命令

dfs -ls -R /; //hive中執行hdfs命令

拷貝表

create table mydb.employee2 as select * from myhive.employee; // 複製表

內部表和外部表

內部表

託管表（內部表），hive默認創建的表都是託管表，hive控制其數據的生命週期。刪除託管表時，元數據和數據均被刪除。

create table if not exists 
mydb.employee_inner(
eid int, 
name String, 
salary String, 
destination String) 
COMMENT 'Employee details' 
ROW FORMAT DELIMITED 
FIELDS TERMINATED BY '\t' 
LINES TERMINATED BY '\n' 
STORED AS TEXTFILE;

drop table mydb.employee_inner;

外部表

hive 控制元數據，刪除外部表時，數據不被刪除

create external table if not exists 
myhive.employee_outer(
eid int, 
name String, 
salary String, 
destination String) 
COMMENT 'Employee details' 
ROW FORMAT DELIMITED 
FIELDS TERMINATED BY '\t' 
LINES TERMINATED BY '\n' 
STORED AS TEXTFILE;

分區表

建表

partitioned by

create table if not exists 
test2(
eid int, 
name String, 
salary String, 
destination String) 
partitioned by(country string,state string) 
ROW FORMAT DELIMITED 
FIELDS TERMINATED BY '\t' 
LINES TERMINATED BY '\n' 
STORED AS TEXTFILE;

描述分區表

載入數據到指定分區

load data local inpath '/root/text.csv' 
into table mydb.test2 
partition(country='china',state='shanxi');

查詢數據

select * from mydb.test2 where country='china' and state='shanxi';

查看 hdfs

分區表的查詢模式: strict/nostrict

默認是非嚴格,嚴格模式要求查詢的時候必須帶分區

set hive.mapred.mode=strict

查看有哪些分區

show partitions mydb.test2;
show partitions mydb.test2 partition(state='shanxi');

增加分區

注意不能增加不存在的分區列

alter table mydb.test2 add partition(country='china',state='henan');

alter table mydb.test2 add 
partition(country='china',state='hubei')  
partition(country='china',state='hebei')  
partition(country='china',state='shandong');

修改表名

alter table mydb.test2 rename to mydb.test;

移動分區的存儲位置

alter table mydb.test partition(country='china',state='hubei') 
set location '/user/hive/warehouse/myhive.db/test/country=china/state=hubei1'; 

驗證一下：

load data local inpath '/root/text.csv' 
into table mydb.test
partition(country='china',state='shanxi');

dfs -ls /user/hive/warehouse/mydb.db/test/country=china/state=hubei;

dfs -ls /user/hive/warehouse/myhive.db/test/country=china;

複製表

create table if not exists
mydb.test3(
eid int,
name String,
salary String,
destination String)
partitioned by(country string,state string)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY '\t'
LINES TERMINATED BY '\n'
STORED AS TEXTFILE;

insert into mydb.test3 
partition(country='china',state='hubei')
select eid,name,salary,destination 
from mydb.test
where country='china' and state='henan';