[
https://issues.apache.org/jira/browse/DRILL-6605?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16545759#comment-16545759
]
Robert Hou commented on DRILL-6605:
-----------------------------------
We have three versions of this query running at sf100.
Drill using hive reader scanning text files passes
Drill using hive reader scanning parquet files does not return any rows
Drill directly scanning views on parquet files timed out on the last run
Interesting, running Hive 1.2 has similar results.
Hive 1.2 returns 100 rows when scanning text files.
Hive 1.2 does not return any rows when scanning parquet files.
The parquet files were created a year ago, so they are old files.
> TPCDS-84 Query does not return any rows
> ---------------------------------------
>
> Key: DRILL-6605
> URL: https://issues.apache.org/jira/browse/DRILL-6605
> Project: Apache Drill
> Issue Type: Bug
> Components: Execution - Relational Operators
> Reporter: Robert Hou
> Assignee: Vitalii Diravka
> Priority: Blocker
> Fix For: 1.14.0
>
>
> Query is:
> Advanced/tpcds/tpcds_sf100/hive/parquet/query84.sql
> This uses the hive parquet reader.
> {code:sql}
> SELECT c_customer_id AS customer_id,
> c_last_name
> || ', '
> || c_first_name AS customername
> FROM customer,
> customer_address,
> customer_demographics,
> household_demographics,
> income_band,
> store_returns
> WHERE ca_city = 'Green Acres'
> AND c_current_addr_sk = ca_address_sk
> AND ib_lower_bound >= 54986
> AND ib_upper_bound <= 54986 + 50000
> AND ib_income_band_sk = hd_income_band_sk
> AND cd_demo_sk = c_current_cdemo_sk
> AND hd_demo_sk = c_current_hdemo_sk
> AND sr_cdemo_sk = cd_demo_sk
> ORDER BY c_customer_id
> LIMIT 100
> {code}
> This query should return 100 rows
> commit id is:
> 1.14.0-SNAPSHOT a77fd142d86dd5648cda8866b8ff3af39c7b6b11
> DRILL-6516: EMIT support in streaming agg 11.07.2018 @ 18:40:03 PDT
> Unknown 12.07.2018 @ 01:50:37 PDT
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
