Question

我正在尝试在Scala中编写Pig UDF（使用Eclipse）。我在java构建路径中添加了pig.jar作为库，它似乎解决了下面的2个导入：

import org.apache.pig.EvalFunc
import org.apache.pig.data.Tuple

然而我得到2个我无法解决的错误：

org.apache.pig.EvalFunc [T]没有构造函数
value get不是org.apache.pig.data.Tuple的成员（虽然我确定Tuple有get方法）

以下是完整代码：

package datesUDFs
import org.apache.pig.EvalFunc
import org.apache.pig.data.Tuple
class getYear extends EvalFunc {
  val extractDate = """^(\d\d\d\d)-\d\d-\d\d \d\d:\d\d:\d\d""".r
  def isDate(dtString: String): Boolean = extractDate.findFirstIn(dtString).nonEmpty

  override def exec(input: Tuple): Int = input.get(0) match {
    case dtString: String =>
      if (!isDate(dtString)) throw new IllegalArgumentException("Invalid date string!")
      else (for (extractDate(year) <- extractDate.findFirstIn(dtString)) yield year).head.toInt
    case _ => throw new IllegalArgumentException("Invalid function call!")
  }
}

有人可以帮我解决这个问题吗？

提前致谢!!!

Answer 1

除了必须指定EvalFunc类型参数外，您的代码对我来说也很合适。

package datesUDFs
import org.apache.pig.EvalFunc
import org.apache.pig.data.Tuple
class getYear extends EvalFunc[Int] { // This is the only line I changed.
  val extractDate = """^(\d\d\d\d)-\d\d-\d\d \d\d:\d\d:\d\d""".r
  def isDate(dtString: String): Boolean = extractDate.findFirstIn(dtString).nonEmpty

  override def exec(input: Tuple): Int = input.get(0) match {
    case dtString: String =>
      if (!isDate(dtString)) throw new IllegalArgumentException("Invalid date string!")
      else (for (extractDate(year) <- extractDate.findFirstIn(dtString)) yield year).head.toInt
    case _ => throw new IllegalArgumentException("Invalid function call!")
  }
}

看看它是否有帮助，有时ScalaIDE会抱怨错误的事情。

Answer 2

解决了！我将 hadoop-common-2.2.0.jar 和 commons-logging-1.1.3.jar 添加到我的java构建路径中，问题得到了解决。

如何在Scala中编写Pig UDF

2 个答案: