hadoop - Pass environment variables to Hive Transform or MapReduce -

i trying pass custom environment variable executable (my-mapper.script in example below) used in hive transform eg:

select    transform(x, y, z)    using 'my-mapper.script' (    select       x, y, z          table )

i know in hadoop streaming can achieved using

-cmdenv example_dir=/home/example/dictionaries/

but not know how in hive transform/mapreduce.

any ideas?

you can wrap script simple 2 line bash script setup environment. e.g

#!/bin/sh export foo=boo my-mapper.script

and use script in query

using 'wrapper.sh'

my-mapper.script see foo (with value "boo") in environment.

Brande