Liu,

You don’t require to build a table, but if you want to  you can. 

You can directly query from the json (ex.   /drill/json/filename.json ) file 
under 

hadoop fs -ls /drill/json/filename.json

using drill

select * from mydfs.json.`filename.json`;

Thanks
Sudhakar Thota


On Dec 30, 2014, at 7:08 PM, LIU Xiaobing <[email protected]> wrote:

> Hi all,
> I am a new user of Apache Drill. Now have set up one distributed hadoop
> server with zookeeper and hbase run in. Then I deployed one clustered
> environment according the wiki.
> Now I have registered one storage plugin instance whose name is mydfs like
> this:
> 
> {
>  "type": "file",
>  "enabled": true,
>  "connection": "hdfs://192.168.41.243:9000/",
>  "workspaces": {
>    "json": {
>      "location": "/drill/json",
>      "writable": true,
>      "defaultInputFormat": null
>    }
>  },
>  "formats": {
>    "json": {
>      "type": "json"
>    }
>  }
> }
> 
> Then should I create one external table pointed to reside hdfs directory
> /drill/json like hive? I havn't found relative
> information about this.
> By the way, I also found that there isn't any default workspace whose name
> is default of file or hive in the section Workspaces.
> Is there anything i missing?
> 
> -- 
> Best Regards
> LIU Xiaobing 刘小兵

Reply via email to