[
https://issues.apache.org/jira/browse/FLINK-25565?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17655739#comment-17655739
]
Sergii Mikhtoniuk commented on FLINK-25565:
-------------------------------------------
Running into this issue too when porting code from {{ParquetRowInputFormat}} in
Flink 1.13 to {{ParquetColumnarRowInputFormat}} in 1.16.
This is a major regression, as {{ParquetRowInputFormat}} were able to read
{{int64}} timestamps just fine.
We use Flink to write data to Parquet via Avro and in 1.13 all timestamps were
written as {{int64}}. So we are now in situation where Flink cannot read the
data that it generated itself.
I'm also confused by these changes because {{int96}} seems to be deprecated in
both Avro and Parquet, so I cannot even force any writer libraries to produce
timestamps in this format any more.
Seeing PR not merged for a year doesn't look reassuring.
> Write and Read Parquet INT64 Timestamp
> --------------------------------------
>
> Key: FLINK-25565
> URL: https://issues.apache.org/jira/browse/FLINK-25565
> Project: Flink
> Issue Type: Sub-task
> Components: Formats (JSON, Avro, Parquet, ORC, SequenceFile)
> Affects Versions: 1.12.0, 1.15.0
> Reporter: Bo Cui
> Priority: Major
> Labels: pull-request-available
>
> Flink cannot read parquet files that contain INT64 Timestamp generated by
> Spark
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
