hive 入门教程，
http://abloz.com
2012.6.14
下载
[zhouhh@Hadoop48 ~]$ wget http://labs.renren.com/apache-mirror/hive/hive-0.9.0/hive-0.9.0.tar.gz
配置
[zhouhh@Hadoop48 ~]$ tar zxvf hive-0.9.0.tar.gz
[zhouhh@Hadoop48 ~]$ cd hive-0.9.0
[zhouhh@Hadoop48 hive-0.9.0]$ export HIVE_HOME=`pwd`
[zhouhh@Hadoop48 hive-0.9.0]$ echo $HIVE_HOME
/home/zhouhh/hive-0.9.0
[zhouhh@Hadoop48 ~]$ vi .bashrc
export HADOOP_HOME=/home/zhouhh/hadoop-1.0.3
export HADOOP_HOME_WARN_SUPPRESS=1
export HBASE_HOME=/home/zhouhh/hbase-0.94.0
export HIVE_HOME=/home/zhouhh/hive-0.9.0
export PATH=$PATH:/sbin/:$HADOOP_HOME/bin:$HBASE_HOME/bin:$HIVE_HOME/bin
[zhouhh@Hadoop48 ~]$ source .bashrc
源码编译：
$ svn co http://svn.apache.org/repos/asf/hive/trunk hive
$ cd hive
$ ant clean package
$ cd build/dist
$ ls
README.txt
bin/ (all the shell scripts)
lib/ (required jar files)
conf/ (configuration files)
examples/ (sample input and query files)
hdfs的/tmp和/user/hive/warehouse必须创建，并且具有组可写权限。
[zhouhh@Hadoop48 ~]$ hadoop fs -ls /
Found 4 items
drwxr-xr-x – zhouhh supergroup 0 2012-06-12 19:15 /hbase
drwxr-xr-x – zhouhh supergroup 0 2012-06-12 12:55 /home
drwxr-xr-x – zhouhh supergroup 0 2012-06-12 12:54 /tmp
drwxr-xr-x – zhouhh supergroup 0 2012-06-11 15:40 /user
[zhouhh@Hadoop48 ~]$ hadoop fs -ls /user
Found 1 items
drwxr-xr-x – zhouhh supergroup 0 2012-06-12 19:04 /user/zhouhh
[zhouhh@Hadoop48 ~]$ hadoop fs -mkdir /user/hive/warehouse
[zhouhh@Hadoop48 ~]$ hadoop fs -chmod g+w /tmp
[zhouhh@Hadoop48 ~]$ hadoop fs -chmod g+w /user/hive/warehouse
[zhouhh@Hadoop48 ~]$ hadoop fs -ls /user/hive/
Found 1 items
drwxrwxr-x – zhouhh supergroup 0 2012-06-14 17:39 /user/hive/warehouse
[zhouhh@Hadoop48 conf]$ pwd
/home/zhouhh/hive-0.9.0/conf
[zhouhh@Hadoop48 conf]$ ls
hive-default.xml.template hive-exec-log4j.properties.template
hive-env.sh.template hive-log4j.properties.template
[zhouhh@Hadoop48 hive-0.9.0]$ ./bin/hive
Exception in thread “main” java.lang.NoClassDefFoundError: org/apache/hadoop/hive/conf/HiveConf
找不到类？
原来，HADOOP_CLASSPATH配置不应该直接覆盖，应将HADOOP_CLASSPATH原来的内容写在前面
export HADOOP_CLASSPATH=$HADOOP_CLASSPATH: …(其他内容)
启动使用
[zhouhh@Hadoop48 hive-0.9.0]$ hive
hive> CREATE TABLE poke(id INT,name STRING);
OK
Time taken: 6.513 seconds
hive> show tables;
OK
poke
Time taken: 0.231 seconds
hive> describe poke;
OK
id int
name string
Time taken: 0.039 seconds
hive> alter table poke add columns(other string comment ‘this is a new column’);
OK
Time taken: 0.237 seconds
hive> describe poke;
OK
id int
name string
other string this is a new column
Time taken: 0.037 seconds
ALTER TABLE [table name] RENAME TO [other name];
DROP TABLE [table name];
[zhouhh@Hadoop48 files]$ pwd
/home/zhouhh/hive-0.9.0/examples/files
[zhouhh@Hadoop48 files]$ head kv1.txt
57^Aval_57
302^Aval_302
205^Aval_205
149^Aval_149
438^Aval_438
345^Aval_345
129^Aval_129
170^Aval_170
…
数据用Ctrl+a分隔的，打印时分隔符看不见。
hive> load data local inpath ‘./examples/files/kv1.txt’ overwrite into table poke;
Copying data from file:/home/zhouhh/hive-0.9.0/examples/files/kv1.txt
Copying file: file:/home/zhouhh/hive-0.9.0/examples/files/kv1.txt
Loading data to table default.poke
Deleted hdfs://Hadoop48:54310/user/hive/warehouse/poke
OK
Time taken: 1.358 seconds
如果是分布式环境，则将命令的local改为hdfs。overwrite 关键字表示覆盖表，会将原来的数据清除。如果没有该关键字，则是追加。
而且，导入数据时不会有数据类型的校验。如果是hdfs，根目录是 hive-default.xml的hive.metastore.warehouse.dir参数指定，建表前应先创建该目录。
hive> select * from poke;
OK
238 val_238 NULL
86 val_86 NULL
311 val_311 NULL
27 val_27 NULL
165 val_165 NULL
…
参考：
https://cwiki.apache.org/confluence/display/Hive/GettingStarted