FileMerger and Unit Test

scala/src/main/scala/ch29_heap_solutions/FileMerger.scala

@@ -0,0 +1,46 @@
+package ch29_heap_solutions
+
+import java.io.{BufferedWriter, File, FileWriter}
+
+import scala.collection.mutable
+import scala.io.Source
+import scala.util.control.Breaks._
+
+object FileMerger {
+
+  /**
+    * each given file has sorted String as content, we need to merge them together
+    *
+    * @param smallFiles - small files with sorted content
+    * @return merged file
+    */
+  def mergeFiles(smallFiles: List[File]): File = {
+    //init output file
+    val output = File.createTempFile("merged-file", ".txt")
+    val writer = new BufferedWriter(new FileWriter(output))
+    //init small top heap
+    val priorityQueue = new mutable.PriorityQueue[Tuple2[Char, Source]]()(Ordering.by((_: (Char, Source))._1).reverse)
+    val sources = smallFiles.toArray.map(smallFile => Source.fromFile(smallFile))
+    //init fill the priority queue from each file
+    sources.foreach(source => priorityQueue.enqueue(Tuple2(source.next(), source)))
+
+    breakable {
+      while (true) {
+        val next = priorityQueue.dequeue()
+        val output: Char = next._1
+        val source = next._2
+        writer.append(output)
+        if (source.hasNext) {
+          priorityQueue.enqueue(Tuple2(source.next(), source))
+        }
+        //determine the end of merge
+        if (sources.forall(!_.hasNext)) {
+          break
+        }
+      }
+    }
+
+    writer.close()
+    output
+  }
+}

scala/src/test/scala/ch29_heap_solutions/FileMergerTest.scala

@@ -0,0 +1,52 @@
+package ch29_heap_solutions
+
+import java.io.{BufferedWriter, File, FileWriter}
+
+import org.scalatest.{FlatSpec, Matchers}
+
+import scala.io.Source
+import scala.util.Random
+
+class FileMergerTest extends FlatSpec with Matchers {
+
+  behavior of "FileMergerTest"
+
+  it should "mergeFiles" in {
+    val num = 10
+    val contentCount = 10
+    val random = Random.alphanumeric
+    val files = new Array[File](num)
+    for (i <- Range(0, num)) {
+      val file = File.createTempFile(i + "-small", ".txt")
+      files(i) = file
+      val writer = new BufferedWriter(new FileWriter(file))
+      val content = random.take((i + 1) * contentCount).toArray.slice(i * contentCount, (i + 1) * contentCount)
+
+      writer.write(content.sorted)
+      writer.flush()
+      writer.close()
+    }
+    println("small files below")
+    files.foreach(printFile)
+
+    val mergedFile = FileMerger.mergeFiles(files.toList)
+
+    val raw = Source.fromFile(mergedFile).toArray
+    raw should equal(raw.sorted)
+
+    println("")
+    println("merged file below")
+    printFile(mergedFile)
+
+    //clean up
+    files.foreach(_.delete())
+    mergedFile.delete()
+
+  }
+
+  def printFile(file: File): Unit = {
+    val source = Source.fromFile(file)
+    source.getLines().foreach(println)
+  }
+
+}