You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "ASF GitHub Bot (Jira)" <ji...@apache.org> on 2022/12/22 18:24:00 UTC
[jira] [Work logged] (HIVE-26885) Iceberg: Parquet Vectorized V2 reads fails with NPE
[ https://issues.apache.org/jira/browse/HIVE-26885?focusedWorklogId=835381&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-835381 ]
ASF GitHub Bot logged work on HIVE-26885:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 22/Dec/22 18:23
Start Date: 22/Dec/22 18:23
Worklog Time Spent: 10m
Work Description: ayushtkn opened a new pull request, #3892:
URL: https://github.com/apache/hive/pull/3892
### What changes were proposed in this pull request?
Fix NPE while reading Iceberg V2 Parquet V2 table in case of empty batches.
### Why are the changes needed?
NPE while reading Parquet V2 table
### Does this PR introduce _any_ user-facing change?
Yep, Query passes without NPE
### How was this patch tested?
UT, On top on HIVE-26884
Issue Time Tracking
-------------------
Worklog Id: (was: 835381)
Remaining Estimate: 0h
Time Spent: 10m
> Iceberg: Parquet Vectorized V2 reads fails with NPE
> ---------------------------------------------------
>
> Key: HIVE-26885
> URL: https://issues.apache.org/jira/browse/HIVE-26885
> Project: Hive
> Issue Type: Bug
> Reporter: Ayush Saxena
> Assignee: Ayush Saxena
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> In case the Iceberg Parquet table lands up having an empty batch, in that case while fetching the row number, used for filtering leads to NPE.
> The row number to block mapping is only done if the parquetSplit isn't null, so in that case, here:
> {code:java}
> if (parquetInputSplit != null) {
> initialize(parquetInputSplit, conf);
> } {code}
> row numbers aren't initialised, so we should skip fetching the row numbers later
--
This message was sent by Atlassian Jira
(v8.20.10#820010)