declaring a List in HiveQL for UDF

2014-04-10 Thread Ng, Raymond
Hi all I'm trying to find out if its possible to pass in a List in a SELECT statement to be used by the "evaluate" method of a UDF using ListObjectInspector on DeferredObject array? I've followed some instructions to build a function which extends GenericUDF (code snippet on "evaluate" method

Hive-exec package JDBC issue

2014-04-10 Thread Valluri, Sathish
Hi all, We are trying to use Hive 0.12 JDBC driver for connecting to remote Hive Server and execute queries. We found a strange issue with this driver that it has dependency on hive-exec package and hive-exec has internally packaged the com.google.common classes into this.(please check the h

HiveServer2 http mode?

2014-04-10 Thread Adam Faris
The "Setting Up HiveServer2" wiki page mentions that HiveServer2 is providing a “http mode" in 0.13. Is “http mode” going to be a rest API or is it encapsulating thrift/jdbc connections inside http traffic? - Thanks, Adam

Re: declaring a List in HiveQL for UDF

2014-04-10 Thread Szehon Ho
I might be wrong, but it seems like it wont be useful, as there is no 'list' Hive data type. Valid complex types are listed here: https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Types#LanguageManualTypes-ComplexTypes As for complex literals, say if you want to test a udf that oper

Re: HiveServer2 http mode?

2014-04-10 Thread Vaibhav Gumashta
Adam, It is doing the latter for now - encapsulating thrift payloads inside http calls. Thanks, --Vaibhav On Thu, Apr 10, 2014 at 1:33 PM, Adam Faris wrote: > The "Setting Up HiveServer2" wiki page mentions that HiveServer2 is > providing a "http mode" in 0.13. Is "http mode" going to be a r

hive query to select top 10 product of each subcategory and select most recent product info

2014-04-10 Thread Mohit Durgapal
I have a hive table partitioned by dates. It contains ecomm data in the format siteid,sitecatid,catid,subcatgid,pid,pname,pprice,pmrp,pdesc What I need to do is to run a query on table above in hive for top 10 products(count wise) in each sub category. What adds a bit more complexity is that