>
> 0: jdbc:drill:zk=local> ALTER SYSTEM SET
> `store.json.read_numbers_as_double` = true;
>
> *+-------+---------------------------------------------+**| ** ok  ** | **
>                 summary                  *
> * |*
> *+-------+---------------------------------------------+**| *true * | 
> *store.json.read_numbers_as_double
> updated.
> * |*
> *+-------+---------------------------------------------+*1 row selected
> (0.075 seconds)
>


> 0: jdbc:drill:zk=local> select t.properties.size from
> dfs.`/Users/khahn/drill/apache-drill-1.1.0-SNAPSHOT/stefan.json` t;
>
> *+---------+**| **EXPR$0 *
> * |*
> *+---------+**| *1
> * |**| *2
> * |*
> *+---------+*2 rows selected (0.096 seconds)


http://drill.apache.org/docs/json-data-model/ has more examples than the
doc page you might be using.

Kristine Hahn
Sr. Technical Writer
415-497-8107 @krishahn


On Sun, Jul 12, 2015 at 1:41 PM, Stefán Baxter <[email protected]>
wrote:

> Hi,
>
> I'm continuing my Drill discovery and trying to build queries against some
> of the data we have.
>
> Our analytics data is a flat structure and I have had no issues working
> with that (apart from parsing dates and dates being stored a binary (more
> on this later))
>
> Now I'm just trying to discover if we can use drill for more generic JSON
> processing but nested data structures are giving me hard times.
>
> Can someone please tell me how to query the following structure for type,
> properties.size
> [
>
>   {
>     "timestamp": "2015-03-31 14:40:53.276",
>     "action": "some.category.action",
>     "origin": "www.some.domain",
>     "properties": {
>       "task_type": "Diagnostic",
>       "size": "1"
>     }
>   },
>   {
>     "timestamp": "2015-03-31 14:40:53.277",
>     "action": "some.category.action",
>     "origin": "www.some.domain",
>     "properties": {
>       "task_type": "Diagnostic",
>       "size": 2,
>       "color": "blue"
>     }
>   }
> ]
>
> * this data is factitious and is purely created for these tests.
>
> These entries have no unique ids (as the donuts document example).
>
> Regards,
>  -Stefan
>

Reply via email to