[
https://issues.apache.org/jira/browse/FLINK-21978?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17309437#comment-17309437
]
Leonard Xu edited comment on FLINK-21978 at 3/26/21, 1:49 PM:
--------------------------------------------------------------
[~twalthr], Snowflake supports cast conversion between INT/BIGINT and
TIMESTAMP_LTZ, both INT value and BIGINT value are treated as seconds, and
Oracle doesn't support the cast conversion between INT/BIGINT and TIMESTAMP
WITH LOCAL TIME ZONE. The conservative approach looks better base on this.
was (Author: leonard xu):
[~twalthr], Snowflake supports cast conversion between INT/BIGINT and
TIMESTAMP, both INT value and BIGINT value are treated as seconds, and Oracle
doesn't support the cast conversion between INT/BIGINT and TIMESTAMP. The
conservative approach looks better base on this.
> Disable cast conversion between Numeric type and TIMESTAMP_LTZ type
> -------------------------------------------------------------------
>
> Key: FLINK-21978
> URL: https://issues.apache.org/jira/browse/FLINK-21978
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / Planner
> Affects Versions: 1.13.0
> Reporter: Leonard Xu
> Assignee: Leonard Xu
> Priority: Major
> Labels: pull-request-available
>
> Currently we has supported the cast conversion between Numeric type and
> TIMESTAMP_LTZ type, we suppose the numeric value e.g `Long type 1000L` as
> epoch seconds and then cast to TIMESTAMP_LTZ, but the java.lang.Long is a
> conversion class of `LocalZonedTimestampType` and treats as milliseconds.
> To avoid the inconsistency, we should disable it and encourage user to use
> `TO_TIMESTAMP_LTZ(numeric, precisoon)` function.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
