Oh, one issue worth raising... if the macros are inside a jar, it is really cool that we can reference UDFs in that jar from the macro. No extra loading needed.
Jacob: would you be interested in contributing the varaha TF-IDF UDF to DataFu? On Thu, Feb 20, 2014 at 8:37 PM, Russell Jurney <russell.jur...@gmail.com>wrote: > Actually, this one by Jacob Perkins is better than mine: > https://github.com/thedatachef/varaha/blob/master/macros/nlp/tfidf.pig > > I rely on default_parallel with macros. I don't see another way if they > are inside a jar. We could make sure the macro source itself has high > visibility for customization/pasting to tune PARALLEL. > > > On Thu, Feb 20, 2014 at 6:47 PM, Sam Shah <shah...@umich.edu> wrote: > >> Can you paste your TFIDF macro? How do you handle parallel statements? >> >> >> On Thu, Feb 20, 2014 at 6:36 PM, Russell Jurney <russell.jur...@gmail.com >> >wrote: >> >> > I would like to add macros to DataFu. I have a TFIDF macro and a couple >> > others I'd like to contribute. >> > >> > What do people think? Any issues that need to be figured out? >> > >> > Russ >> > >> > >> > -- >> > Russell Jurney twitter.com/rjurney russell.jur...@gmail.com >> > datasyndrome.com >> > >> > > > > -- > Russell Jurney twitter.com/rjurney russell.jur...@gmail.com datasyndrome. > com > -- Russell Jurney twitter.com/rjurney russell.jur...@gmail.com datasyndrome.com
