Is the current plan to make HCatalog the only way be which one can read
and/or write ORC files? It seems like there are cases where it would be
nice to use ORC files outside of Pig, Hive, Map Reduce, perhaps as an
alternative to Parquet.
Thanks,
Andrew
On 9/27/13 12:36 AM, "Thejas Nair" wrot
you can use hcatalog input and output formats to read/write ORC from mapreduce.
http://hive.apache.org/docs/hcat_r0.5.0/inputoutput.html
On Thu, Sep 26, 2013 at 9:34 PM, 陈涛 wrote:
> i am trying to write orcfile(snappy) instead of sequencefile(lzo) in
> mapreduce,but OrcStruct is not public incl
i am trying to write orcfile(snappy) instead of sequencefile(lzo) in
mapreduce,but OrcStruct is not public include getFieldValue method:
class OrcStruct implements Writable {
private Object[] fields;
OrcStruct(int children) {
fields = new Object[children];
}
Object getFieldValue(in
