spark - 简记

map阶段写内存 - 归并排序
SortShufleWriter.write
sorter.writePartitionedFile(blockId, tmp)
collection.destructiveSortedWritablePartitionedIterator(comparator)
destructiveSortedIterator(comparator)
new Sorter(new KVArraySortDataFormat[K, AnyRef]).sort(data, 0, newIndex, keyComparator)
在这里插入图片描述

猜你喜欢

转载自blog.csdn.net/qq_29573903/article/details/88385809