Hi, I am developing an Open Source OLAP application called "Cubulus". The code is at http://sourceforge.net/projects/cubulus/ , a brief presentation material at http://cubulus.sourceforge.net/ , and an online demo at: http://alxtoth.webfactional.com
It would be interresting to use Pig instead of relational databases as backend. The question is: can Pig scripts work is such manner that the file is loaded only once, and then subsequent web requests process over and over the same file? This becomes relevant if the data file is large, and there is one datafile to process (or few datafiles). In fact, is repated loading a problem at all :-) ? -Alex
