hive 文件系统学习实例

Hive存储是基于hadoop hdfs文件系统的，通过默认内嵌的Derby 数据库或外部数据库系统（如mysql)组织元数据访问，下面就通过实际案例描述其存储过程。

1， 在hive 中创建表，然后把外部csv文件导入其中（外部文件为Batting.csv, 内部表为temp_batting):

hive>create table temp_batting(col_value STRING);

hive> show tables;
OK
temp_batting
...

hive>LOAD DATAINPATH'hive/data/Batting.csv' OVERWRITE INTO TABLE temp_batting;

2, 查看外部mysql数据库，可以看到新创建的temp_batting表：

mysql> use hive;
Reading table information for completion of table and column names
You can turn off this feature to get a quicker startup with -A

mysql> select * from TBLS;
+--------+-------------+-------+------------------+-------+-----------+-------+--------------+----------------+--------------------+--------------------+
| TBL_ID | CREATE_TIME | DB_ID | LAST_ACCESS_TIME | OWNER | RETENTION | SD_ID |TBL_NAME | TBL_TYPE | VIEW_EXPANDED_TEXT |VIEW_ORIGINAL_TEXT |
+--------+-------------+-------+------------------+-------+-----------+-------+--------------+----------------+--------------------+----------
| 66 | 1432707070 | 1 | 0 | root | 0 | 66 | temp_batting | MANAGED_TABLE |NULL | NULL |
| |
+--------+-------------+-------+------------------+-------+-----------+-------+--------------+----------------+--------------------+----------

...

查看其在hdfs上存储路径：

mysql> select * from SDS;
+-------+-------+--------------------------------------------------+---------------+---------------------------+--------------------------------------------------------+-------------+------------------------------------------------------------+----------+
| SD_ID | CD_ID | INPUT_FORMAT |IS_COMPRESSED | IS_STOREDASSUBDIRECTORIES | LOCATION | NUM_BUCKETS |OUTPUT_FORMAT | SERDE_ID |
+-------+-------+--------------------------------------------------+---------------+---------------------------+--------------------------------
| 66 | 71 | org.apache.hadoop.mapred.TextInputFormat | | |hdfs://localhost:9000/user/hive/warehouse/temp_batting | -1 |org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat | 66 |

可以看到是:
hdfs://localhost:9000/user/hive/warehouse/temp_batting

3，到hadoop 的hdfs文件系统中查看这个表路径：

[root@lr rli]# hadoop dfs -ls /user/hive/warehouse
DEPRECATED: Use of this script to execute hdfs command is deprecated.
Instead use the hdfs command for it.
...

drwxr-xr-x - root supergroup 02015-05-27 14:16 /user/hive/warehouse/temp_batting
...

[root@lr rli]# hadoop dfs -ls/user/hive/warehouse/temp_batting

DEPRECATED: Use of this script to execute hdfs command isdeprecated.
Instead use the hdfs command for it.


Found 1 items
-rwxr-xr-x 1 root supergroup 6398990 2015-05-2714:02 /user/hive/warehouse/temp_batting/Batting.csv

可以看到其文件大小及内容。

结论：

Hive通过关联数据库系统记录文件的存储路径，属性等，实际数据存在hdfs系统中，当通过select等操作生成相应的map/reduce进程进一步数据分析处理。