[
https://issues.apache.org/jira/browse/PARQUET-1297?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16473085#comment-16473085
]
ASF GitHub Bot commented on PARQUET-1297:
-----------------------------------------
masayuki038 opened a new pull request #477: PARQUET-1297: SchemaConverter
should not convert from Timestamp(TimeUnit.SECOND) and
Timestamp(TimeUnit.NANOSECOND) of Arrow
URL: https://github.com/apache/parquet-mr/pull/477
Arrow's 'Timestamp' definition is below:
```json
{
"name" : "timestamp",
"unit" : "SECOND|MILLISECOND|MICROSECOND|NANOSECOND"
}
```
http://arrow.apache.org/docs/metadata.html
But Parquet only supports 'TIMESTAMP_MILLIS' and 'TIMESTAMP_MICROS'.
https://github.com/Apache/parquet-format/blob/master/LogicalTypes.md
Therefore SchemaConverter should not convert from Timestamp(TimeUnit.SECOND)
and Timestamp(TimeUnit.NANOSECOND) of Arrow to Parquet.
Related:
- https://issues.apache.org/jira/browse/PARQUET-1285
- https://github.com/apache/parquet-mr/pull/469
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> [Java] SchemaConverter should not convert from Timestamp(TimeUnit.SECOND) and
> Timestamp(TimeUnit.NANOSECOND) of Arrow
> ---------------------------------------------------------------------------------------------------------------------
>
> Key: PARQUET-1297
> URL: https://issues.apache.org/jira/browse/PARQUET-1297
> Project: Parquet
> Issue Type: Bug
> Components: parquet-mr
> Reporter: Masayuki Takahashi
> Assignee: Masayuki Takahashi
> Priority: Minor
> Fix For: 1.10.0
>
>
> Arrow's 'Timestamp' definition is below:
> {code:java}
> {
> "name" : "timestamp",
> "unit" : "SECOND|MILLISECOND|MICROSECOND|NANOSECOND"
> }
> {code}
> [http://arrow.apache.org/docs/metadata.html]
> But Parquet only supports 'TIMESTAMP_MILLIS' and 'TIMESTAMP_MICROS'.
> [https://github.com/Apache/parquet-format/blob/master/LogicalTypes.md]
> Therefore SchemaConverter should not convert from Timestamp(TimeUnit.SECOND)
> and Timestamp(TimeUnit.NANOSECOND) of Arrow to Parquet.
> Related:
> https://issues.apache.org/jira/browse/PARQUET-1285
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
