[
https://issues.apache.org/jira/browse/DRILL-5763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16150728#comment-16150728
]
ASF GitHub Bot commented on DRILL-5763:
---------------------------------------
Github user amansinha100 commented on a diff in the pull request:
https://github.com/apache/drill/pull/931#discussion_r136601989
--- Diff:
contrib/format-maprdb/src/main/java/org/apache/drill/exec/store/mapr/db/MapRDBSubScan.java
---
@@ -71,14 +71,32 @@ public MapRDBSubScan(String userName,
MapRDBFormatPlugin formatPlugin,
this.tableType = tableType;
}
+
+ @JsonProperty("formatPluginConfig")
--- End diff --
There are trade-offs with serializing/deserializing more than the minimum
needed. One is the extra amount of data being sent as part of the plan
fragment. Another, somewhat more problematic is that the serialization brings
in dependent classes and if the latter are not serialized properly then query
fails. Any thoughts about these ? I understand the motivation for the change
to fix DRILL-5714 and subsequently this Jira. Just want to explore it further
to see if the downsides can be avoided.
> Fix NPE during MapRDBSubScan deserialization
> --------------------------------------------
>
> Key: DRILL-5763
> URL: https://issues.apache.org/jira/browse/DRILL-5763
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - MapRDB
> Affects Versions: 1.12.0
> Reporter: Arina Ielchiieva
> Assignee: Arina Ielchiieva
> Labels: ready-to-commit
> Fix For: 1.12.0
>
>
> When querying mapr-db json table, if query has more than one major fragments,
> query execution fails since not all properties are defined MapRDBSubScan and
> thus deserialization fails.
> This is the regression from DRILL-5714.
> The root cause of the issue is that {{formatPluginConfig}} and
> {{storageConfig}} were not indicated as json properties and thus where not
> serialized. During deserialization they were set as null which caused NPE.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
