danny0405 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1829129236
yes, the mor reader merges the payloads before returning the result set.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
zdl1 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1828955749
> the compaction writer actually knows the record operation when it does the
payload merging
Sorry for the late reply, what if I just `select count(*) from table`? Does
it
danny0405 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1823978838
yeah, not very good way to collect the correct numbers, a good chance if for
compaction, the compaction writer actually knows the record operation when it
does the payload merging.
zdl1 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1823873879
> there is no way to figure out whether a key has been written to an
existing bucket before, except the first file slice, so all the records are
updates.
Thanks for the
danny0405 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1822589574
Yeah, correct.
--
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.
zdl1 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1822541106
There are 2 delta_commit files generated by 2 insertions
danny0405 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1822537964
yes, you are right, actually in bucket index, there is no way to figure out
whether a key has been written to an existing bucket before, except the first
file slice, so all the
zdl1 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1822501418
> The bucket type `UPSERT` and record operation(update/delete) are different
things, an `upsert` buckets just means we write more inc records to existing
data bucket.
Thanks for
danny0405 commented on issue #10156:
URL: https://github.com/apache/hudi/issues/10156#issuecomment-1822468337
The bucket type `UPSERT` and record operation(update/delete) are different
things, an `upsert` buckets just means we write more inc records to existing
data bucket.
--
This is
zdl1 opened a new issue, #10156:
URL: https://github.com/apache/hudi/issues/10156
When inserting a record to a MOR table, trying to assign a bucket for the
record
10 matches
Mail list logo
