Comments on: mapReduce Reduced (& Ported to R) http://opendatagroup.com/2009/09/10/mapreduce-reduced-ported-to-r/ Open Data Group's Home Page and Blog Wed, 14 Jul 2010 22:38:22 -0500 http://wordpress.org/?v=2.8.6 hourly 1 By: Rodrigo Miranda (rodrigocmiranda) 's status on Tuesday, 17-Nov-09 11:31:53 UTC - Identi.ca http://opendatagroup.com/2009/09/10/mapreduce-reduced-ported-to-r/comment-page-1/#comment-39 Rodrigo Miranda (rodrigocmiranda) 's status on Tuesday, 17-Nov-09 11:31:53 UTC - Identi.ca Tue, 17 Nov 2009 11:31:57 +0000 http://blog.opendatagroup.com/?p=77#comment-39 [...] in R: http://blog.opendatagroup.com/2009/09/10/mapreduce-reduced-ported-to-r/ a few seconds ago from [...] [...] in R: http://blog.opendatagroup.com/2009/09/10/mapreduce-reduced-ported-to-r/ a few seconds ago from [...]

]]> By: Adrian http://opendatagroup.com/2009/09/10/mapreduce-reduced-ported-to-r/comment-page-1/#comment-30 Adrian Thu, 10 Sep 2009 16:55:28 +0000 http://blog.opendatagroup.com/?p=77#comment-30 Great idea to implement mapReduce. From what I understand it seems that the data you start with needs to be loaded into R. This is obvious a serious limitation when you use large data sets. In the Google paper, one starts from a set of files and the map part is done in such a way that the produced chunks are small enough to be processed by the workers It would be nice to avoid having to load the initial data into memory. Great idea to implement mapReduce. From what I understand it seems that the data you start with needs to be loaded into R. This is obvious a serious limitation when you use large data sets. In the Google paper, one starts from a set of files and the map part is done in such a way that the produced chunks are small enough to be processed by the workers

It would be nice to avoid having to load the initial data into memory.

]]>