Hi Prajakta,
I get it now. But are there plans to support groupbys of that format,
or would I be able to write a UDF of sorts to do that?
Also, could you please give me an example of accelerating equi-joins
with indices, I haven't been able to get that to work either :(
Regards,
--
Rohan Monga
Hi Rohan
As of now, we support queries with aggregate function count on columns which
are the same as index key column and are used in group-by construct. For
example,
CREATE TABLE tbl(key int, value int);
CREATE INDEX tbl_key_idx ON TABLE tbl(key) AS
'org.apache.hadoop.hive.ql.index.AggregateInde
Hi Prajakta,
Thanks, I was able to generate a plan which was using indices by
following your advise.
However, I want to group on one key and count the others
"select count(id2) from table_t1 group by id1", would that not be possible ?
Also, could you give me an example of using indices to speed up
Hi Rohan
I run your queries on the same version from trunk and did not get any
errors. Not sure why you are getting a semantic analysis error. It would be
good if you could send me a snapshot of the error message.
About the second issue of getting the same plan even if you
set hive.optimize.index
