using JSON-SerDe in Hive tables

问题:

I'm trying JSON-SerDe from below link http://code.google.com/p/hive-json-serde/wiki/GettingStarted.

         CREATE TABLE my_table (field1 string, field2 int, 
                                     field3 string, field4 double)
         ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.JsonSerde' ;

I've added Json-SerDe jar as

          ADD JAR /path-to/hive-json-serde.jar;

And loaded data as

LOAD DATA LOCAL INPATH  '/home/hduser/pradi/Test.json' INTO TABLE my_table;

and it loads data successfully.

But when query data as

Select * from my_table;

I get only one row from table as

data1 100 more data1 123.001

Test.json contains

{"field1":"data1","field2":100,"field3":"more data1","field4":123.001} 

{"field1":"data2","field2":200,"field3":"more data2","field4":123.002} 

{"field1":"data3","field2":300,"field3":"more data3","field4":123.003} 

{"field1":"data4","field2":400,"field3":"more data4","field4":123.004}

Where is the problem? why only one row is coming instead of 4 rows when i query the table. And in /user/hive/warehouse/my_table contains all the 4 rows!!

---

hive> add jar /home/hduser/pradeep/hive-json-serde-0.2.jar;
Added /home/hduser/pradeep/hive-json-serde-0.2.jar to class path
Added resource: /home/hduser/pradeep/hive-json-serde-0.2.jar

hive> CREATE EXTERNAL TABLE my_table (field1 string, field2 int,
>                                 field3 string, field4 double)
> ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.JsonSerde'
> WITH SERDEPROPERTIES (
>   "field1"="$.field1",
>   "field2"="$.field2",
>   "field3"="$.field3",
>   "field4"="$.field4"
> );
OK
Time taken: 0.088 seconds

hive> LOAD DATA LOCAL INPATH  '/home/hduser/pradi/test.json' INTO TABLE my_table;
Copying data from file:/home/hduser/pradi/test.json
Copying file: file:/home/hduser/pradi/test.json
Loading data to table default.my_table
OK
Time taken: 0.426 seconds

hive> select * from my_table;
OK
data1   100     more data1      123.001
Time taken: 0.17 seconds

I've already posted the contents of test.json file. so you can see that query is resulting only one line as

data1   100     more data1      123.001

---

I've changed the json file to employee.json which contains

{ "firstName" : "Mike", "lastName" : "Chepesky", "employeeNumber" : 1840192 }

and changed table also but it showing a null values when i query the table

hive> add jar /home/hduser/pradi/hive-json-serde-0.2.jar;
Added /home/hduser/pradi/hive-json-serde-0.2.jar to class path
Added resource: /home/hduser/pradi/hive-json-serde-0.2.jar

hive> create EXTERNAL table employees_json (firstName string, lastName string,        employeeNumber int )
> ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.JsonSerde';
OK
Time taken: 0.297 seconds


hive> load data local inpath '/home/hduser/pradi/employees.json' into table     employees_json;
Copying data from file:/home/hduser/pradi/employees.json
Copying file: file:/home/hduser/pradi/employees.json
Loading data to table default.employees_json
OK
Time taken: 0.293 seconds


 hive>select * from employees_json;
  OK
  NULL    NULL    NULL
  NULL    NULL    NULL
  NULL    NULL    NULL
  NULL    NULL    NULL
  NULL    NULL    NULL
  NULL    NULL    NULL
Time taken: 0.194 seconds

回答1:

A bit hard to tell what's going on without the logs (see Getting Started) in case of doubt. Just a quick thought - can you try if it works with WITH SERDEPROPERTIESas so:

CREATE EXTERNAL TABLE my_table (field1 string, field2 int, 
                                field3 string, field4 double)
ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.JsonSerde'
WITH SERDEPROPERTIES (
  "field1"="$.field1",
  "field2"="$.field2",
  "field3"="$.field3",
  "field4"="$.field4" 
);

There is also a fork you might want to give a try from ThinkBigAnalytics.

UPDATE: Turns out the input in Test.json is invalid JSON hence the records get collapsed.

See answer https://stackoverflow.com/a/11707993/396567 for further details.

回答2:

1. First of all you have to validate your json file on http://jsonlint.com/ after that make your file as one row per line and remove the [ ]. the comma at the end of the line is mandatory.

   [{"field1":"data1","field2":100,"field3":"more data1","field4":123.001}, {"field1":"data2","field2":200,"field3":"more data2","field4":123.002}, {"field1":"data3","field2":300,"field3":"more data3","field4":123.003}, {"field1":"data4","field2":400,"field3":"more data4","field4":123.004}]

2. In my test I added hive-json-serde-0.2.jar from hadoop cluster , I think hive-json-serde-0.1.jar should be ok.

   ADD JAR hive-json-serde-0.2.jar;

3. Create your table

   CREATE TABLE my_table (field1 string, field2 int, field3 string, field4 double) ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.JsonSerde' ;

4. Load your Json data file ,here I load it from hadoop cluster not from local

   LOAD DATA INPATH 'Test2.json' INTO TABLE my_table;

My test

回答3:

for json parsing based on cwiki/confluence we need follow some steps

1. need to download hive-hcatalog-core.jar

2. hive> add jar /path/hive-hcatalog-core.jar

3. create table tablename(colname1 datatype,.....) row formatserde'org.apache.hive.hcatalog.data.JsonSerDe' stored as ORCFILE;

4. colname in creating table and colname in test.json must be same if not it will show null values Hope it wil helpfull