Rahul Challapalli created DRILL-1288:
----------------------------------------
Summary: With json data drill is not returning all the columns in
the query involving a join
Key: DRILL-1288
URL: https://issues.apache.org/jira/browse/DRILL-1288
Project: Apache Drill
Issue Type: Bug
Components: Storage - JSON
Reporter: Rahul Challapalli
Priority: Critical
git.commit.id.abbrev
In the below case drill does not return the 3rd column
{code}
select
s.s_suppkey,
s.s_nationkey,
n.n_name
from
supplier s,
nation n
where
s.s_nationkey = n.n_nationkey limit 1;
+------------+-------------+
| s_suppkey | s_nationkey |
+------------+-------------+
| 1 | 17 |
+------------+-------------+
{code}
In the below case drill does not return the second column
{code}
select
n.n_name,
s.s_suppkey
from
supplier s,
nation n
where
s.s_nationkey = n.n_nationkey limit 1;
+------------+
| n_name |
+------------+
| PERU |
+------------+
{code}
Looks like drill is only returning the columns from table/file of the first
column. If the first column is from nation, drill returns all subsequent
columns from nation but does not return anything from supplier.
However if use a function on columns from the second table, drill does return
everything normally
{code}
select
s.s_suppkey,
char_length(n.n_name) as name_length
from
supplier s,
nation n
where
s.s_nationkey = n.n_nationkey limit 1;
+------------+-------------+
| s_suppkey | name_length |
+------------+-------------+
| 1 | 4 |
+------------+-------------+
{code}
This issue is affecting TPCH queries 10, 7, 3 on top of JSON data.
Data used is provided as an attachement
--
This message was sent by Atlassian JIRA
(v6.2#6252)
