java - Hadoop Wordcount Example with multiple files and kill mappers in it

Question

I have 3 input files : file1.txt , file2.txt, file3.txt. file1.txt contains some error data. When i submit these files in MapReduce job 3 mappers will be created. Now i need to kill the mapper which contains the error file called file1.txt. After killing the mapper i want the error file to be discarded from the mapper and store in it hdfs. How can i kill the mapper and discard the file??
And if i kill the mapper the remaining mapper should complete their process and entire job should be complete . Please help

Answer 1

这可能不是一种选择方法。将元组 (K, V) 存储在映射内存中，如果不满足无效条件，则在最后发出它们。