[ https://issues.apache.org/jira/browse/FLINK-30612 ]
guoxin deleted comment on FLINK-30612:
--------------------------------
was (Author: guoxin):
the same problem exists with inner-class MapConverter and ObjectArrayConverter
in DataFormatConverters.
the latest version still has this problem. obviously.
when i define a implementation used to query set and list data in redis, i get
messy data.
> ObjectArrayConverter is not thread-safe，raised by reuseArray、reuseWriter
> ------------------------------------------------------------------------
>
> Key: FLINK-30612
> URL: https://issues.apache.org/jira/browse/FLINK-30612
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / API, Table SQL / Runtime
> Affects Versions: 1.12.2
> Reporter: guoxin
> Priority: Major
>
> data clutter in multi-threaded environment.
> by design?
> or , how do i use it in multi-thread env?
> {code:java}
> package org.apache.flink.table.data.util;
> public class DataFormatConverters{
> ....
> public static final class ObjectArrayConverter<T> extends
> DataFormatConverter<ArrayData, T[]>{
> ....
> // it's not thread-safe
> private transient BinaryArrayData reuseArray;
> private transient BinaryArrayWriter reuseWriter;
> ....
> private ArrayData toBinaryArray(T[] value){
> // not thread-safe
> }
> ....
> }
> ....
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
