Scala正则表达式的java.lang.NullPointerException

Question

I am having a bizarre behavior in my scala project when I try to run it from sbt.

The line that throws a java.lang.NullPointerException is this:

id <- urlReg.findFirstIn(line).map(_.split("/").lift(3))

It is inside a for-comprehension:

for {
  line <- Source.fromFile(fileName).getLines()
  id <- urlReg.findFirstIn(line).map(_.split("/").lift(3))
  dateString <- dateReg.findFirstMatchIn(line).map(_.group(1))
} yield {...}

When I test this line at a scala console, I see that findFirstIn returns an Option[String] , so it should return a None or Some[String] and never throw a java.lang.NullPointerException .

This code ran for awhile without throwing the exception and then started throwing the exception.

urlReg is val urlReg = "[GET,POST,DELETE,PUT]\\\\s{1}[\\\\w/]+".r

A sample line of input is

val sampleLine = """10.10.6.90 - - 15/Aug/2016:23:59:20 -0500 "GET /ecf8427e/b443dc7f/71f28176/174ef735/1dd4d421 HTTP/1.0" 200 - "-" "-" 7 "10.10.1.231, 10.10.6.90" -"""

I expect there is some incompatibility in my setup. I am running:

• MacOS Sierra 10.12.1.
• sbt 0.13.13
• scala 2.12.1
• java 1.8.0_112

Answer 1

I also believe that it is due to some Option[String] type that are returned by 2 of the function in your script.

Would you consider using a map/filter combination like the one below:

val faultyLine = """10.10.6.90 - - 15/Aug/2016:23:59:20 -0500 "GET /ecf8427eb443dc7f71f28176174ef7351dd4d421 HTTP/1.0" 200 - "-" "-" 7 "10.10.1.231, 10.10.6.90" -"""
val line= """10.10.6.90 - - 15/Aug/2016:23:59:20 -0500 "GET /ecf8427e/b443dc7f/71f28176/174ef735/1dd4d421 HTTP/1.0" 200 - "-" "-" 7 "10.10.1.231, 10.10.6.90" -"""

val lines = List(line,line,faultyLine)

lines.map(line => (urlReg.findFirstIn(line).map(_.split("/").lift(3))).match {
case Some(Some(a)) => a 
case _ => null
}).filter( _ != null )

You should be able to replace the map/filter by a flatMap but I was not successful at it. That should simplify it and get rid of the ugly nulls in there.

Good luck!