Check TextInputFormat. You could override it to achieve that.
________________________________ From: Kunsheng Chen <ke...@yahoo.com> To: core-user@hadoop.apache.org Sent: Saturday, May 23, 2009 5:04:50 PM Subject: Could hadoop do word count in all files under two-level sub folders? Hello everyone, I referred to the hadoop tutorial online and found that wordcount example, it seems to me that all files have to be under a certain folder to make it work. I am not sure whether that workcount example could work for multiple subfolders. For example, if the input folder is 'input' , and I have two subfolders 'input/input1', 'input/input2', will it work if I only tell the program folder 'input' ? Or I have to program a little bit myself for that? I know it is a simple quesiton that could be tried myself, but the thing is that I haven't decided whether to use hadoop for my project and not yet installed it. So any simple answer or idea is well apprecaited, Thanks a lot! -Kun