[
https://issues.apache.org/jira/browse/NIFI-8610?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17348982#comment-17348982
]
Dominik Przybysz commented on NIFI-8610:
----------------------------------------
Hi [~pgyori] ,
I checked your PR and you are fixing a problem with reading AVROs using another
reader than in mine PR, because convertAvroToORC uses [embedded
schema|https://github.com/apache/nifi/blob/main/nifi-nar-bundles/nifi-standard-services/nifi-record-serialization-services-bundle/nifi-record-serialization-services/src/main/java/org/apache/nifi/avro/AvroReaderWithEmbeddedSchema.java].
In my opinion the whole problem is with reading AVROs with reusing previous
generic record not only in the situation you described. Maybe reusing generic
record should be removed from all places in AVRO readers (and from its API) or
such reuse should be fixed.
To conclude: we both fixed different problems with the same root.
> Invalid bigdecimal values in ORC when multiple avro records provided
> --------------------------------------------------------------------
>
> Key: NIFI-8610
> URL: https://issues.apache.org/jira/browse/NIFI-8610
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework
> Affects Versions: 1.13.2
> Reporter: Dominik Przybysz
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Processor: convertAvroToORC
> Input: flow file contains a lot of avro records with decimal values as bytes
> with logical type decimal
> Output: ORC file with different values for decimal values
> Reason: AvroReader used in processor reuses previous generic record when
> fetching next one
> Solution: Don't reuse generic records when reading records from flow file
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
