I am new to Hadoop. I have to create a prefix tree from the data set(which is divided across nodes) and use the output of the reducer in phase one as a input in phase two and work on the prefix tree and output the key value pairs for phase two . The tree is constructed only once during phase one from the local dataset. Is it possible that in the phase two the created prefix tree will be in memory that is dont want to create it again. Thank you -- Sameer Awasekar "To love is nothing, to be loved is something but to love and to be loved by the one you love,that is EVERYTHING." - Unknown
