When comparing HParser with the integrated Data Transformation that comes with Big Data Management, HParser was found to perform better. Integrated Data Transformation was about ~2.57X slower in some internal tests. Informatica recommends trying out and comparing the performance of HParser and integrated Data Transformation before arriving at a conclusion.

If you use Data Transformation, Informatica recommends using a splittable input format if available and design the Data Processor transformation's streaming service to process a batch of records using the count property of the streamer.

For more information, see the following topics in the

Informatica PowerExchange for HDFS User Guide

:

See Case Study: HParser vs. Integrated Data Transformation