Dear Mr Kumar, I also recently thinking about it. From my humble experience in QlikView side, I presumed such:
- QlikView requires data to be ready (=processable): whatever we do in Hadoop side, with Map + Reduce, QlikView would require visualisable dataset.
- To create visualisable dataset, we would require either a programs or tools, working closely on Hadoop. Other tools might be more suitable for hypothesis-driven research (using such languages like R).
In the meantime, I have found some info in Cloudera about connector. We could explore it.
So this might lead to my impression that QlikView would be only useful for last-mile of dev/research team: to present the data to the business users. For the rest, it seems we need to dive into the world of Java, R, .... etc.
Recently I encountered an expert on Hadoop who mentioned that the adoptions of Hadoop is stagnant in 2014-15 due to the massive requirement of specialists in addition to the existing faculties (if it were to be implemented purely from open source). I would be happy hearing your recent experiences in Hadoop projects in relation to the QlikView.
