The "uname="$._u" is the correct form. We also hacked on this SerDe at Think Big Analytics. I don't know if you'll see an improvement though.
https://github.com/thinkbiganalytics/hive-json-serde I wonder if there's a problem handling the leading underscore? Also, I know it's just an example, but in case it was taken from a real situation, the dates in your example are for January. dean On Mon, Feb 18, 2013 at 6:43 AM, Chunky Gupta <chunky.gu...@vizury.com>wrote: > Hi, > > I have data in s3 bucket, which is in json format and is a zip file. I > have added this jar file in hive console :- > > http://code.google.com/p/hive-json-serde/downloads/detail?name=hive-json-serde-0.2.jar&can=2&q= > > I tried the following steps to create table and load data :- > > 1. CREATE EXTERNAL TABLE table_test ( uname STRING ) PARTITIONED BY (dt > STRING ) ROW FORMAT SERDE "org.apache.hadoop.hive.contrib.serde2.JsonSerde" > WITH SERDEPROPERTIES ( "uname"="$._u" ) LOCATION > 's3://BUCKET_NAME/test_data/' > > I tried this also :- > > CREATE EXTERNAL TABLE table_test ( uname STRING ) PARTITIONED BY (dt > STRING ) ROW FORMAT SERDE "org.apache.hadoop.hive.contrib.serde2.JsonSerde" > WITH SERDEPROPERTIES ( "uname"="_u" ) LOCATION > 's3://BUCKET_NAME/test_data/' > > > > 2. alter table table_test add partition (dt='13Feb2012') location > 's3n://BUCKET_NAME/test_data/13Feb2012'; > > and json file is like this :- > ------------------------------------- > {"_u":"test_name1","_ts":"2012-01-13","_ip":"IP1"} > {"_u":"test_name2","_ts":"2012-01-13","_ip":"IP2"} > {"_u":"test_name3","_ts":"2012-01-13","_ip":"IP3"} > > > When I query :- > select uname from table_test; > > Output :- > NULL 13Feb2012 > NULL 13Feb2012 > NULL 13Feb2012 > > > Please help me and let me know how to add json data in a table. > > Thanks, > Chunky. > -- *Dean Wampler, Ph.D.* thinkbiganalytics.com +1-312-339-1330
