[GitHub] [hudi] garyli1019 commented on a change in pull request #1702: Bootstrap datasource changes

[GitBox <gi...@apache.org>]

garyli1019 commented on a change in pull request #1702:
URL: https://github.com/apache/hudi/pull/1702#discussion_r449933225



##########
File path: hudi-spark/src/main/scala/org/apache/hudi/HudiBootstrapRDD.scala
##########
@@ -0,0 +1,131 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *      http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hudi
+
+import org.apache.spark.{Partition, TaskContext}
+import org.apache.spark.rdd.RDD
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.execution.datasources.PartitionedFile
+import org.apache.spark.sql.types.StructType
+import org.apache.spark.sql.vectorized.ColumnarBatch
+
+class HudiBootstrapRDD(@transient spark: SparkSession,
+                       dataReadFunction: PartitionedFile => Iterator[Any],
+                       skeletonReadFunction: PartitionedFile => Iterator[Any],
+                       regularReadFunction: PartitionedFile => Iterator[Any],
+                       dataSchema: StructType,
+                       skeletonSchema: StructType,
+                       requiredColumns: Array[String],
+                       tableState: HudiBootstrapTableState)
+  extends RDD[InternalRow](spark.sparkContext, Nil) {
+
+  override def compute(split: Partition, context: TaskContext): Iterator[InternalRow] = {
+    val bootstrapPartition = split.asInstanceOf[HudiBootstrapPartition]
+
+    if (log.isDebugEnabled) {
+      if (bootstrapPartition.split.skeletonFile.isDefined) {
+        logDebug("Got Split => Index: " + bootstrapPartition.index + ", Data File: "
+          + bootstrapPartition.split.dataFile.filePath + ", Skeleton File: "
+          + bootstrapPartition.split.skeletonFile.get.filePath)
+      } else {
+        logDebug("Got Split => Index: " + bootstrapPartition.index + ", Data File: "
+          + bootstrapPartition.split.dataFile.filePath)
+      }
+    }
+
+    var partitionedFileIterator: Iterator[InternalRow] = null
+
+    if (bootstrapPartition.split.skeletonFile.isDefined) {
+      // It is a bootstrap split. Check both skeleton and data files.
+      if (dataSchema.isEmpty) {
+        // No data column to fetch, hence fetch only from skeleton file
+        partitionedFileIterator = read(bootstrapPartition.split.skeletonFile.get,  skeletonReadFunction)
+      } else if (skeletonSchema.isEmpty) {
+        // No metadata column to fetch, hence fetch only from data file
+        partitionedFileIterator = read(bootstrapPartition.split.dataFile, dataReadFunction)
+      } else {
+        // Fetch from both data and skeleton file, and merge
+        val dataFileIterator = read(bootstrapPartition.split.dataFile, dataReadFunction)
+        val skeletonFileIterator = read(bootstrapPartition.split.skeletonFile.get, skeletonReadFunction)
+        partitionedFileIterator = merge(skeletonFileIterator, dataFileIterator)
+      }
+    } else {
+      partitionedFileIterator = read(bootstrapPartition.split.dataFile, regularReadFunction)
+    }
+    partitionedFileIterator
+  }
+
+  def merge(skeletonFileIterator: Iterator[InternalRow], dataFileIterator: Iterator[InternalRow])

Review comment:
       I think this approach is better than extending the `FileFormat`. Ultimately, we can have a `HudiRDD` to handle all the file loading and merging(bootstrap files, parquet, orc, logs). `Union` will trigger shuffle and grouping files on the driver then use different `FileFormat` to read is not as clean as this approach. 
   I will add the `MOR` stuff on top of this PR after this merged.




----------------------------------------------------------------
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:
users@infra.apache.org