Github user dongjoon-hyun commented on the issue:
https://github.com/apache/spark/pull/13765
Hi, @cloud-fan , @liancheng , @gatorsmile .
We can imagine some data sciences environment having many predefined
system-wide tables or datasets. But, here, I simplify that into an extremely
two line example.
`dsView1` is one of the system-wide predefined table. The number of
partition of `dsView` is optimized for the common usage for the performance (or
for preventing skewness.)
Data scientist can not modify that Dataset because it's a shared and
predefined one.
However, Data scientist want to save that data into a single file. (You
can see this kind of pattern in current Spark MLlib code, too.)
```
scala> val dsView1 = spark.range(8).repartition(8)
scala> dsView1.repartition(1).write.json("/tmp/jsonsinglefile")
```
You can see that the obvious repetition of `repartition`s.
How do you think about this virtual scenario?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
