Thank you for the quick response. I'm using composition to implement
Giraph's VertexInputFormat/VertexOutputFormat, so it's fine that
mapreduce is used. Has HCatInputFormat/HCatOutputFormat been tested
successfully with Hive's latest releases? I'd love to see someone's
example code if it exists to load data from HCatInputFormat...
Avery
On 11/18/11 12:49 PM, David Capwell wrote:
Currently hcat only supports the mapreduce package and not the mapred package
that giraph uses.
Also in order to use hcat your table properties must define "hcat.isd" and
"hcat.osd" which stand for InputStorageDriver and OutputStorageDriver. These are the
interfaces used in hcat to read/write data.
On Nov 18, 2011, at 12:36:04PM, Avery Ching wrote:
Hi,
I was interested in trying out HCatalog to access Hive tables for Apache
Giraph (basically a map-only job). I tried out some code from
HCatMapReduceTest, but had some issues getting it to work. Is there
some documentation on how to use HCatInputFormat properly with Hive tables?
Thanks,
Avery
