Extend TextInputFormat and change the behavior any way you like. You can look at org.apache.hadoop.mapred.KeyValueTextInputFormat for an example.
On 9/19/07 6:58 AM, "贺皓(He Hao)" <[EMAIL PROTECTED]> wrote: > Hi Everyone, > I'm a newer to hadoop. I want to write a mapreduce program to implement the > inverted index. My question is which input format should I use? It seems that > the TextInputFormat's key is the offset of each line in the document. How can > I get the document name(as the document id) that the map function processes? > > Thanks! > > He Hao > > > --------------------------------- > @yahoo.cn 新域名、无限量，快来抢注！
