Github user HyukjinKwon commented on a diff in the pull request:
https://github.com/apache/spark/pull/17395#discussion_r107822893
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileFormatWriter.scala
---
@@ -292,7 +292,10 @@ object FileFormatWriter extends Logging {
override def execute(iter: Iterator[InternalRow]): Set[String] = {
var fileCounter = 0
var recordsInFile: Long = 0L
- newOutputWriter(fileCounter)
+ // Skip the empty partition to avoid creating a mass of 'empty'
files.
+ if (iter.hasNext) {
+ newOutputWriter(fileCounter)
--- End diff --
Yes, I was thinking in that way. I remember I tried several tries at that
time but failed to make a good fix, and could not have some time to work on
that further.
Another problem is, it might be a datasource-specific issue because, for
example, ORC does not write out empty df. For example,
```
scala> spark.range(100).filter("id > 100").write.orc("/tmp/abc1")
scala> spark.read.orc("/tmp/abc1").show()
org.apache.spark.sql.AnalysisException: Unable to infer schema for ORC. It
must be specified manually.;
at
org.apache.spark.sql.execution.datasources.DataSource$$anonfun$8.apply(DataSource.scala:182)
at
org.apache.spark.sql.execution.datasources.DataSource$$anonfun$8.apply(DataSource.scala:182)
```
This issue is described in
https://issues.apache.org/jira/browse/SPARK-15474.
FWIW, I happened to see https://issues.apache.org/jira/browse/SPARK-15693
around that time and I kind of felt we may be able to consolidate this issue
with it although it is a rough idea.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
