On Mon, Apr 9, 2012 at 5:29 PM, Deepak Nettem <deepaknet...@gmail.com> wrote: > Hi Stan, > > Just out of curiosity, care to explain the use case a bit? >
Very simply: lots of reasonably small files which I can't control, i.e., changing block size is not an option. Note that this is not an issue in pig or hive, both of which come with their own implementation of CombineFileInputFormat.