Thanks, Raghu, it works. However, I have to manually copy the map/reduce script to each node in the cluster in exactly the same path. This is not required if I use ADD FILE from the cli. Any chance of having such a feature from the JDBC API?
Saurabh. On Mon, Aug 10, 2009 at 9:51 PM, Raghu Murthy <[email protected]> wrote: > I mean, don't use the 'ADD' statement. Just directly specify the path in > your SELECT statement. > > SELECT TRANSFORM(a,b) USING '/path/to/script' AS c,d > FROM table > > > On 8/10/09 1:58 AM, "Saurabh Nanda" <[email protected]> wrote: > > > > >> As a work around, can you copy the file locally on the server and use > the > >> complete path (local to the server) in the query? > > > > The file is available locally to the machine running "./bin/hive > --service > > hiveserver" Do you want me to try using the full file path? However, the > error > > message doesn't say "file not found". It says "cannot recognize input > 'add'" > > > > Saurabh. > > -- http://nandz.blogspot.com http://foodieforlife.blogspot.com
