jackylk commented on a change in pull request #3798:
URL: https://github.com/apache/carbondata/pull/3798#discussion_r446609807



##########
File path: 
integration/spark/src/main/scala/org/apache/carbondata/api/CarbonStore.scala
##########
@@ -63,13 +68,101 @@ object CarbonStore {
     }
 
     if (limit.isDefined) {
-      val lim = Integer.parseInt(limit.get)
-      segmentsMetadataDetails.slice(0, lim)
+      segmentsMetadataDetails.slice(0, limit.get)
     } else {
       segmentsMetadataDetails
     }
   }
 
+  /**
+   * Read stage files and return input files
+   */
+  def readStages(
+      tablePath: String,
+      configuration: Configuration): Seq[StageInput] = {
+    val stageFiles = listStageFiles(
+      CarbonTablePath.getStageDir(tablePath), configuration)
+    var output = Collections.synchronizedList(new util.ArrayList[StageInput]())
+    output.addAll(readStageInput(stageFiles._1,
+      StageInput.StageStatus.Unload).asJavaCollection)
+    output.addAll(readStageInput(stageFiles._2,
+      StageInput.StageStatus.Loading).asJavaCollection)
+    Collections.sort(output, new Comparator[StageInput]() {
+      def compare(stageInput1: StageInput, stageInput2: StageInput): Int = {
+        (stageInput2.getCreateTime - stageInput1.getCreateTime).intValue()
+      }
+    })
+    output.asScala
+  }
+
+  /**
+   * Read stage files and return input files
+   */
+  def readStageInput(
+      stageFiles: Seq[CarbonFile],
+      status: StageInput.StageStatus): Seq[StageInput] = {
+    val gson = new Gson()
+    val output = Collections.synchronizedList(new util.ArrayList[StageInput]())
+    stageFiles.map { stage =>
+      val filePath = stage.getAbsolutePath
+      val stream = FileFactory.getDataInputStream(filePath)
+      try {
+        val stageInput = gson.fromJson(new InputStreamReader(stream), 
classOf[StageInput])
+        stageInput.setCreateTime(stage.getLastModifiedTime)
+        stageInput.setStatus(status)
+        output.add(stageInput)
+      } finally {
+        stream.close()
+      }
+    }
+    output.asScala
+  }
+
+  /*
+   * Collect all stage files and matched success files.
+   * A stage file without success file will not be collected
+   */
+  def listStageFiles(
+        loadDetailsDir: String,
+        hadoopConf: Configuration): (Array[CarbonFile], Array[CarbonFile]) = {
+    val dir = FileFactory.getCarbonFile(loadDetailsDir, hadoopConf)
+    if (dir.exists()) {
+      var allFiles = dir.listFiles()
+      val successFiles = allFiles.filter { file =>
+        file.getName.endsWith(CarbonTablePath.SUCCESS_FILE_SUBFIX)
+      }.map { file =>
+        (file.getName.substring(0, file.getName.indexOf(".")), file)
+      }.toMap
+      val loadingFiles = allFiles.filter { file =>
+        file.getName.endsWith(CarbonTablePath.LOADING_FILE_SUBFIX)
+      }.map { file =>
+        (file.getName.substring(0, file.getName.indexOf(".")), file)
+      }.toMap
+
+      allFiles = allFiles.filter { file =>
+        !file.getName.endsWith(CarbonTablePath.SUCCESS_FILE_SUBFIX)
+      }.filter { file =>
+        !file.getName.endsWith(CarbonTablePath.LOADING_FILE_SUBFIX)
+      }
+
+      val unloadedFiles = allFiles.filter { file =>
+        successFiles.contains(file.getName)

Review comment:
       maybe you can first construct the last part of the file by "file.getName 
+ CarbonTablePath.LOADING_FILE_SUBFIX", then can still use endsWith. I think it 
can be a little faster




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Reply via email to