[
https://issues.apache.org/jira/browse/SPARK-53368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18043545#comment-18043545
]
Uroš Bojanić commented on SPARK-53368:
--------------------------------------
[~dongjoon] Please move this ticket back to
https://issues.apache.org/jira/browse/SPARK-51162 (Time SPIP umbrella).
> Consider supporting Parquet TIME data with isAdjustedToUTC=true
> ---------------------------------------------------------------
>
> Key: SPARK-53368
> URL: https://issues.apache.org/jira/browse/SPARK-53368
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Affects Versions: 4.1.0
> Reporter: Sarah Gilmore
> Priority: Minor
>
> It would be nice if the Parquet datasource supported reading Parquet {{TIME}}
> types with {{isAdjustedToUTC=true.}} Currently, the Spark Parquet reader
> errors if the {{isAdjustedToUTC}} parameter is not {{false.}}
> Some Parquet writers, such as Arrow's Parquet writer, write {{TIME}} data
> with {{isAdjustedToUTC=true}} by default because of the [Parquet
> compatibility
> guidelines|https://github.com/apache/parquet-format/blob/master/LogicalTypes.md#deprecated-time-convertedtype]
> regarding the deprecation of the {{ConvertedType}} {{{}TIME_MICROS{}}}. For
> example, Arrow's Parquet writer currently writes TIME data with
> {{isAdjustedToUTC=true}} by default even though Arrow {{TIME}} types are
> timezone-agnostic.
> NOTE: There is [movement|https://github.com/apache/arrow/pull/47316] in the
> Arrow community to change the {{isAdjustedToUTC}} to {{false}} by default.
> However, there are still many pre-existing Parquet files that were generated
> by Arrow that contain {{TIME}} data that cannot be imported by the Spark
> Parquet reader.
> See [this [email protected] email
> thread|https://lists.apache.org/thread/cxy9obkf74177w4tsxwq6oqjx6t95ksk] for
> more details.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
