[
https://issues.apache.org/jira/browse/HIVE-17961?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vihang Karajgaonkar updated HIVE-17961:
---------------------------------------
Attachment: HIVE-17961.03.patch
The qtest {{vectorization_parquet_projection.q}} works for me when I run
locally. Looking at the logs I see this error.
{noformat}
javax.jdo.JDODataStoreException: Error executing SQL query "select
"PARTITIONS"."PART_ID" from "PARTITIONS" inner join "TBLS" on
"PARTITIONS"."TBL_ID" = "TBLS"."TBL_ID" and "TBLS"."TBL_NAME" = ? inner
join "DBS" on "TBLS"."DB_ID" = "DBS"."DB_ID" and "DBS"."NAME" = ? inner
join "PARTITION_KEY_VALS" "FILTER0" on "FILTER0"."PART_ID" =
"PARTITIONS"."PART_ID" and "FILTER0"."INTEGER_IDX" = 0 where (((case when
"FILTER0"."PART_KEY_VAL" <> ? and "TBLS"."TBL_NAME" = ? and "DBS"."NAME" = ?
and "FILTER0"."PART_ID" = "PARTITIONS"."PART_ID" and "FILTER0"."INTEGER_IDX" =
0 then cast("FILTER0"."PART_KEY_VAL" as decimal(21,0)) else null end) = ?))"
{noformat}
The directSQL failed when partition key is not a string and Hive creates
partition "__HIVE_DEFAULT_PARTITION__". This could be larger problem and not
related to this JIRA. Uploaded version 3 of the patch which modifies the test
to use String type for partition key instead of int.
> NPE during initialization of VectorizedParquetRecordReader when input split
> is null
> -----------------------------------------------------------------------------------
>
> Key: HIVE-17961
> URL: https://issues.apache.org/jira/browse/HIVE-17961
> Project: Hive
> Issue Type: Sub-task
> Components: Hive
> Affects Versions: 3.0.0, 2.4.0
> Reporter: Vihang Karajgaonkar
> Assignee: Vihang Karajgaonkar
> Attachments: HIVE-17961.01.patch, HIVE-17961.02.patch,
> HIVE-17961.03.patch
>
>
> HIVE-16465 introduces the regression which causes a NPE during initialize of
> the vectorized reader when input split is null. This was already fixed in
> HIVE-15718 but got exposed again we refactored for HIVE-16465. We should also
> add a test case to catch such regressions in the future.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
