Anton Gozhiy created DRILL-7429:
-----------------------------------
Summary: Wrong column order when selecting complex data using Hive
storage plugin.
Key: DRILL-7429
URL: https://issues.apache.org/jira/browse/DRILL-7429
Project: Apache Drill
Issue Type: Bug
Affects Versions: 1.16.0
Reporter: Anton Gozhiy
Attachments: customer_complex.zip
*Data:*
customer_complex.zip attached
*Query:*
{code:sql}
select t3.a, t3.b from (select t2.a, t2.a.o_lineitems[1].l_part.p_name b from
(select t1.c_orders[0] a from hive.customer_complex t1) t2) t3 limit 1
{code}
*Expected result:*
Column order: a, b
*Actual result:*
Column order: b, a
*Physical plan:*
{noformat}
00-00 Screen
00-01 Project(a=[ROW($0, $1, $2, $3, $4, $5, $6, $7)], b=[$8])
00-02 Project(a=[ITEM($0, 0).o_orderstatus], a1=[ITEM($0,
0).o_totalprice], a2=[ITEM($0, 0).o_orderdate], a3=[ITEM($0,
0).o_orderpriority], a4=[ITEM($0, 0).o_clerk], a5=[ITEM($0, 0).o_shippriority],
a6=[ITEM($0, 0).o_comment], a7=[ITEM($0, 0).o_lineitems],
b=[ITEM(ITEM(ITEM(ITEM($0, 0).o_lineitems, 1), 'l_part'), 'p_name')])
00-03 Project(c_orders=[$0])
00-04 SelectionVectorRemover
00-05 Limit(fetch=[10])
00-06 Scan(table=[[hive, customer_complex]],
groupscan=[HiveDrillNativeParquetScan [entries=[ReadEntryWithPath
[path=/drill/customer_complex/000000_0]], numFiles=1, numRowGroups=1,
columns=[`c_orders`[0].`o_orderstatus`, `c_orders`[0].`o_totalprice`,
`c_orders`[0].`o_orderdate`, `c_orders`[0].`o_orderpriority`,
`c_orders`[0].`o_clerk`, `c_orders`[0].`o_shippriority`,
`c_orders`[0].`o_comment`, `c_orders`[0].`o_lineitems`,
`c_orders`[0].`o_lineitems`[1].`l_part`.`p_name`]]])
{noformat}
*Note:* Reproduced with both Hive and Native readers. Non-reproducible with
Parquet reader.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
