flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From GitBox <...@apache.org>
Subject [GitHub] [flink] JingsongLi commented on issue #10035: [FLINK-14080][table-planner-blink] Introduce Timestamp as internal representation of TIMESTAMP_WITHOUT_TIME_ZONE
Date Thu, 31 Oct 2019 06:36:15 GMT
JingsongLi commented on issue #10035: [FLINK-14080][table-planner-blink] Introduce Timestamp
as internal representation of TIMESTAMP_WITHOUT_TIME_ZONE
URL: https://github.com/apache/flink/pull/10035#issuecomment-548235482
 
 
   > adding these methods long getEpochMilliSecond(int idx), int getNano(int idx), ZoneId
getZoneId(int idx).
   The reason is the same as why decimal not getLong directly.
   It will make CodeGen and computation complex, and if we separate timestamp precision less
than 3 and greater than 3, it will bring two implementations on timestamp type for all functions,
and it is not worth. And with JIT, object will be optimized in many scene. One type should
keep only one presentation.
   
   > it is not an internal representation
   Data format is not only internal use, so the `CONVERSION` set of logical type contains
data format too. It will cause confusion. In Java, the distinction between instant and LocalDateTime
is not about performance, but about their semantics and methods. Its comments may confuse
us.
   If it is just internal use, why not add a class to do some optimize and implement functions
and etc.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

Mime
View raw message