Github user HyukjinKwon commented on the pull request:
https://github.com/apache/spark/pull/12818#issuecomment-216740931
@jbax Cool! Thank you for detailed explanation.
So, this uses OS default newline without `setLineSeparator()`, which is
trimmed
[here](https://github.com/apache/spark/blob/73b56a3c6c5c590219b42884c8bbe88b0a236987/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/csv/CSVParser.scala#L89)
for each row in Spark.
```scala
scala> "foo\n".stripLineEnd
res0: String = foo
scala> "foo\r\n".stripLineEnd
res1: String = foo
```
(FYI, actually, currently Saprk writes line by line but [it opens and
closes `CSVWriter` for each
row](https://github.com/apache/spark/blob/73b56a3c6c5c590219b42884c8bbe88b0a236987/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/csv/CSVParser.scala#L79-L91),
which definitely should be refactored. This uses another Haddop thridparty
library `LineRecordWriter` to actually write the each line parsed to a string
by Univocity. So, I noticed `setLineSeparator()` can be ignorable.)
---
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]
