flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1617403962
ok
--
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.
To unsubscribe,
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1617185826
> That's true. Actually it is even more friendly for Hive query engine too,
just a little late for 0.14.0 release because I'm scared for introducing
protential bug, we can make the first
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1617185611
> That's true. Actually it is even more friendly for Hive query engine too,
just a little late for 0.14.0 release because I'm scared for introducing
protential bug, we can make the first
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1612907539
> The `DeltaCommitWriteHandleFactory` can be tweaked for the purpose, I'm
wondering what's the engine conflicts you are talking about?
sry to reply late
## engine conflicts:
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1612904526
>
sry to reply late
## engine conflicts:
v0.12.2 when spark insert overwrite a partition after flink write the log
files only bucket in this partition,
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1607062700
First align flinkās logic to spark to generate base file on the first write
resolve many engine conflicts and makes logic consistent,
Second, the change will not affect flink's current
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1607048249
> > I'm puzzled with this as spark generate base file first
>
> Because Spark uses BloomFilter index, with wich a base file metadata is
maintained for all the new inserts.
Can
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1607027978
@hudi-bot run azure
--
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.
To
flashJd commented on PR #9048:
URL: https://github.com/apache/hudi/pull/9048#issuecomment-1606692365
Flink generate log files at the first data written on MOR, I'm puzzled with
this as spark generate base file first, I want to know the consideration
@danny0405
--
This is an automated