Bob Rumsby created DRILL-1418:
---------------------------------
Summary: Drill Tutorial join of Hive table and JSON file fails on
Sandbox build
Key: DRILL-1418
URL: https://issues.apache.org/jira/browse/DRILL-1418
Project: Apache Drill
Issue Type: Bug
Affects Versions: 0.5.0
Environment: Using the Drill Sandbox build running 0.5 from Friday,
Sept. 12.
This query worked previously on a demo cluster set up by Andy Pernsteiner with
the following build:
[mapr@ip-172-16-1-60 drill-0.5.0]$ more git.properties
#Generated by Git-Commit-Id-Plugin
#Mon Aug 18 12:53:44 PDT 2014
git.commit.id.abbrev=654c879
The data set required to run this query is on Andy's cluster as well as on the
Sandbox build, but the storage configuration will be a little different.
Reporter: Bob Rumsby
0: jdbc:drill:> use dfs.clicks;
+------------+------------+
| ok | summary |
+------------+------------+
| true | Default schema changed to 'dfs.clicks' |
+------------+------------+
1 row selected (0.121 seconds)
0: jdbc:drill:> select o.cust_id, o.order_total,t.trans_info.prod_id[0] from
hive.orders as o, `clicks/clicks.json` t where o.cust_id=t.user_info.cust_id
and t.trans_info.prod_id[0] is not null and o.order_total > (select
avg(inord.order_total) from hive.orders inord where inord.state = o.state );
Query failed: Failure while running fragment. Schema is currently null. You
must call buildSchema(SelectionVectorMode) before this container can return a
schema. [f4394fe6-d419-4196-ab84-2628f5541a51]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
