Hive Cli

Hive Cli 

= hive启动  =  
*$ hive –f script.q 
*$ hive -e 'SELECT * FROM dummy‘ 
*$ hive -S -e 'SELECT * FROM dummy‘ 
*$ hive -hiveconf hive.root.logger=DEBUG,console 

= set  =  
*hive> SET hive.enforce.bucketing=true;  //设置值 
*hive> SET hive.enforce.bucketing;  // 显示值 
*hive> set -v;  //显示所有的值，包括hadoop的。 
*hive> set; //显示跟基本的hadoop不同的配置，原理就是比对当前的所有选项与基本的配置是否不同，如修改过，已经不同了就打印该值。 

= dfs 命令 =  
dfs 命令可以执行 Shell 环境下的 hadoop fs 的所有命令 
例如： 
列出 HDFS 上的文件： 
hive> dfs -ls /user/hive; 

= add  =  
*ADD { FILE[S] | JAR[S] | ARCHIVE[S] } <filepath1> [<filepath2>]* 
*hive> add jar /tmp/a.jar; 
*hive> add jar /tmp/a.jar /tmp/b.jar; 

= delete  =  
*DELETE { FILE[S] | JAR[S] | ARCHIVE[S] } [<filepath1> <filepath2> ..] 

= list  =  
*LIST { FILE[S] | JAR[S] | ARCHIVE[S] } [<filepath1> <filepath2> ..] 
*hive> list jar; 
*hive> list jars; 

= show  =  
*hive> show functions; 
*hive> show tables; 
*hive> show tables '*tianzhao*'; 
*hive> show partition tablename; 
*hive> show table extended like table_with_partitions partition(ds=101); 

= desc  =  
*hive> desc/describe function length; 
*hive> desc/describe tablename; 
*hive> desc/describe extended tablename;  //显示表的信息 
*hive> desc/describe extended tablename partition(ds=1); //显示partition的信息 

*hive> desc/describe formatted tablename;  //显示表的信息,跟extended相比，显示更友好 

field.delim是分隔符参数

serialization.null.format 用于null值的存储格式设置

= source  =  
*hive> source /home/username/sql/test.sql; 

= !  =  
*hive> !ls;  //ls当前目录 

= quit  =  
hive> quit; 或者  hive> exit; 

= hiveserver  =  
* $hive --service hiveserver 
* $hive --service help 
Usage ./hive <parameters> --service serviceName <service parameters> 
Service List: cli help hiveserver hwi jar lineage metastore rcfilecat start-hive stop-hive 
Parameters parsed: 
  --auxpath : Auxillary jars 
  --config : Hive configuration directory 
  --service : Starts specific service/component. cli is default 
Parameters used: 
  HADOOP_HOME : Hadoop install directory 
  HIVE_OPT : Hive options 
For help on a particular service: 
  ./hive --service serviceName --help 


*$hive --service start-hive 
Starting Hive Thrift Server in Daemon Mode 
starting jar, logging to /home/tianzhao/hive/hadoop-0.19.1/bin/../logs/hadoop-tianzhao-jar-ubuntu.out 
*$hive --service stop-hive 
Stopping Hive Thrift Server in Daemon Mode 
stopping jar 


[http://wiki.apache.org/hadoop/Hive/LanguageManual/Cli 官方的配置Wiki] 

alter table s_spu set TBLPROPERTIES ('EXTERNAL'='TRUE');  //内部表转外部表 

set hive.auto.convert.join=true; 
set hive.exec.mode.local.auto=true; 
set hive.mapred.local.mem = 200; 
set hive.groupby.skewindata=true; 

alter serde  
数据： 
a,b 
c,d 
e,f 
hive> create table delim(key string, value string); 
hive> load data local inpath '/home/tianzhao/Documents/delim' into table delim; 
hive> select * from delim; 
a,b NULL 
c,d NULL 
e,f NULL 
hive> ALTER TABLE delim SET SERDEPROPERTIES ('field.delim' = ','); 
a b 
c d 
e f 

https://cwiki.apache.org/confluence/display/Hive/LanguageManual+DDL#LanguageManualDDL-DropPartitions 
ALTER TABLE page_view DROP PARTITION (dt='2008-08-08', country='us'); 

ALTER TABLE web_log ADD IF NOT EXISTS PARTITION (pt='20120325') LOCATION 'hdfs://localhost:9000/group/log1/2012/20120325'; 
ALTER TABLE web_log PARTITION(pt='20120325') SET FILEFORMAT SequenceFile; 
ALTER TABLE r_table SET SERDEPROPERTIES ('serialization.null.format'=); 

日志在： /tmp/$USER/hive.log  中