Github user reggert commented on the issue:
https://github.com/apache/spark/pull/15899
I don't get why you say that it "doesn't even work in general". Under what
circumstances doesn't it work? I've never run into any problems with it.
The "simple syntactic sugar" allows very clear, concise code to be written
in many cases, and even lets you take advantage of Scala pattern matching for
filtering. For example:
```scala
val strings = sparkContext.parallelize(List("1213,999", "abc", "456,789"))
val NumberPairString = """(\d{1,5}),(\d{1,5})""".r
val numbers = for (NumberPairString(a, b) <- strings; n <- Seq(a, b)) yield
n.toInt
// numbers.collect() yields Array[Int](1213, 999, 456, 789)
```
Without the `for` comprehension, you wind up with this significantly uglier
and somewhat confusing chain of calls:
```scala
val strings = sparkContext.parallelize(List("1213,999", "abc", "456,789"))
val NumberPairString = """(\d{1,5}),(\d{1,5})""".r
val numbers = strings.filter {
case NumberPairString(_, _) => true
case _ => false
}.flatMap{
case NumberString(a, b) => Seq(a, b)
}.map(_.toInt)
// numbers.collect() yields Array[Int](1213, 999, 456, 789)
```
There are alternate ways to write this (e.g., with a single `flatMap` on a
pattern match function that returns either 0 or 2 elements), but none of them
are as clean and concise as the `for` version.
---
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]
