hadoop - MapReduce read file from classpath in Tasks -

i have bundled fat jar file "xxx.txt.gz"

i need reference file inside each yarn container inside each map task.

so if inside jar:

you see xxx.txt.gz*

i trying access file via

file mappingfile = new file(getclass().getclassloader().getresource("xxx.txt.gz").getfile()) 

however, @ run time following error logs task attempts

java.io.filenotfoundexception: file:/local/hadoop/1/yarn/local/usercache/user/appcache/application_1431608807540_0071/filecache/10/job.jar/job.jar!/xxx.txt.gz (no such file or directory) 

in other words, though fat jar had file, thejob.jar not.

how can remedy this?

thanks lot in advance.

there way of accessing file mappers/reducers. hope idea might ideal in mapreduce.

you can use distributed cache option available in mapreduce. way can make hadoop distribute file containers on job's mappers/reducers execute.