日志
Hive创建索引
索引是标准的数据库技术，hive 0.7版本之后支持索引。Hive提供有限的索引功能，这不像传统的关系型数据库那样有“键(key)”的概念，用户可以在某些列上创建索引来加速某些操作，给一个表创建的索引数据被保存在另外的表中。 Hive的索引功能现在还相对较晚，提供的选项还较少。但是，索引被设计为可使用内置的可插拔的java代码来定制，用户可以扩展这个功能来满足自己的需求。 当然不是说有的查询都会受惠于Hive索引。用户可以使用EXPLAIN语法来分析HiveQL语句是否可以使用索引来提升用户查询的性能。像RDBMS中的索引一样，需要评估索引创建的是否合理，毕竟，索引需要更多的磁盘空间，并且创建维护索引也会有一定的代价。 用户必须要权衡从索引得到的好处和代价。
下面说说怎么创建索引：
| 1 | hive> create table user( id int, name string) |
| 2 | > ROW FORMAT DELIMITED |
| 3 | > FIELDS TERMINATED BY '\t' |
| 4 | > STORED AS TEXTFILE; |
2、导入数据：
| 1 | hive> load data local inpath '/export1/tmp/wyp/row.txt' |
| 2 | > overwrite into table user; |
3、创建索引之前测试
| 01 | hive> select * from user where id =500000; |
| 02 | Total MapReduce jobs = 1 |
| 03 | Launching Job 1 out of 1 |
| 04 | Number of reduce tasks is set to 0 since there's no reduce operator |
| 05 | Cannot run job locally: Input Size (= 356888890) is larger than |
| 06 | hive.exec.mode.local.auto.inputbytes.max (= 134217728) |
| 07 | Starting Job = job_1384246387966_0247, Tracking URL = |
| 08 |
| 09 | http://l-datalogm1.data.cn1:9981/proxy/application_1384246387966_0247/ |
| 10 |
| 11 | Kill Command=/home/q/hadoop/bin/hadoop job -kill job_1384246387966_0247 |
| 12 | Hadoop job information for Stage-1: number of mappers:2; number of reducers:0 |
| 13 | 2013-11-13 15:09:53,336 Stage-1 map = 0%, reduce = 0% |
| 14 | 2013-11-13 15:09:59,500 Stage-1 map=50%,reduce=0%, Cumulative CPU 2.0 sec |
| 15 | 2013-11-13 15:10:00,531 Stage-1 map=100%,reduce=0%, Cumulative CPU 5.63 sec |
| 16 | 2013-11-13 15:10:01,560 Stage-1 map=100%,reduce=0%, Cumulative CPU 5.63 sec |
| 17 | MapReduce Total cumulative CPU time: 5 seconds 630 msec |
| 18 | Ended Job = job_1384246387966_0247 |
| 19 | MapReduce Jobs Launched: |
| 20 | Job 0: Map: 2 Cumulative CPU: 5.63 sec |
| 21 | HDFS Read: 361084006 HDFS Write: 357 SUCCESS |
| 22 | Total MapReduce CPU Time Spent: 5 seconds 630 msec |
| 23 | OK |
| 24 | 500000 wyp. |
| 25 | Time taken: 14.107 seconds, Fetched: 1 row(s) |
一共用了14.107s
4、对user创建索引
| 01 | hive> create index user_index on table user(id) |
| 02 | > as 'org.apache.hadoop.hive.ql.index.compact.CompactIndexHandler' |
| 03 | > with deferred rebuild |
| 04 | > IN TABLE user_index_table; |
| 05 | hive> alter index user_index on user rebuild; |
| 06 | hive> select * from user_index_table limit 5; |
| 07 | 0 hdfs://mycluster/user/hive/warehouse/table02/000000_0 [0] |
| 08 | 1 hdfs://mycluster/user/hive/warehouse/table02/000000_0 [352] |
| 09 | 2 hdfs://mycluster/user/hive/warehouse/table02/000000_0 [704] |
| 10 | 3 hdfs://mycluster/user/hive/warehouse/table02/000000_0 [1056] |
| 11 | 4 hdfs://mycluster/user/hive/warehouse/table02/000000_0 [1408] |
| 12 | Time taken: 0.244 seconds, Fetched: 5 row(s) |
这样就对user表创建好了一个索引。
5、对创建索引后的user再进行测试
| 01 | hive> select * from user where id =500000; |
| 02 | Total MapReduce jobs = 1 |
| 03 | Launching Job 1 out of 1 |
| 04 | Number of reduce tasks is set to 0 since there's no reduce operator |
| 05 | Cannot run job locally: Input Size (= 356888890) is larger than |
| 06 | hive.exec.mode.local.auto.inputbytes.max (= 134217728) |
| 07 | Starting Job = job_1384246387966_0247, Tracking URL = |
| 08 |
| 09 | http://l-datalogm1.data.cn1:9981/proxy/application_1384246387966_0247/ |
| 10 |
| 11 | Kill Command=/home/q/hadoop/bin/hadoop job -kill job_1384246387966_0247 |
| 12 | Hadoop job information for Stage-1: number of mappers:2; number of reducers:0 |
| 13 | 2013-11-13 15:23:12,336 Stage-1 map = 0%, reduce = 0% |
| 14 | 2013-11-13 15:23:53,240 Stage-1 map=50%,reduce=0%, Cumulative CPU 2.0 sec |
| 15 | 2013-11-13 15:24:00,253 Stage-1 map=100%,reduce=0%, Cumulative CPU 5.27 sec |
| 16 | 2013-11-13 15:24:01,650 Stage-1 map=100%,reduce=0%, Cumulative CPU 5.27 sec |
| 17 | MapReduce Total cumulative CPU time: 5 seconds 630 msec |
| 18 | Ended Job = job_1384246387966_0247 |
| 19 | MapReduce Jobs Launched: |
| 20 | Job 0: Map: 2 Cumulative CPU: 5.63 sec |
| 21 | HDFS Read: 361084006 HDFS Write: 357 SUCCESS |
| 22 | Total MapReduce CPU Time Spent: 5 seconds 630 msec |
| 23 | OK |
| 24 | 500000 wyp. |
| 25 | Time taken: 13.042 seconds, Fetched: 1 row(s) |
时间用了13.042s这和没有创建索引的效果差不多。
在Hive创建索引还存在bug：如果表格的模式信息来自SerDe，Hive将不能创建索引：
| 01 | hive> CREATE INDEX employees_index |
| 02 | > ON TABLE employees (country) |
| 03 | > AS 'org.apache.hadoop.hive.ql.index.compact.CompactIndexHandler' |
| 04 | > WITH DEFERRED REBUILD |
| 05 | > IDXPROPERTIES ('creator' = 'me','created_at' = 'some_time') |
| 06 | > IN TABLE employees_index_table |
| 07 | > COMMENT 'Employees indexed by country and name.'; |
| 08 | FAILED: Error in metadata: java.lang.RuntimeException: \ |
| 09 | Check the index columns, they should appear in the table being indexed. |
| 10 | FAILED: Execution Error, return code 1 from \ |
| 11 | org.apache.hadoop.hive.ql.exec.DDLTask |
这个bug发生在Hive0.10.0、0.10.1、0.11.0，在Hive0.12.0已经修复了，详情请参见：https://issues.apache.org/jira/browse/HIVE-4251
/2 
