| 1 |
18684
|
mateusz.ko
|
To Build:
|
| 2 |
|
|
|
| 3 |
|
|
mvn clean package
|
| 4 |
|
|
|
| 5 |
|
|
Hive Usage:
|
| 6 |
|
|
|
| 7 |
|
|
add jar avro-json-1.0-SNAPSHOT.jar;
|
| 8 |
|
|
CREATE TABLE doctors (foo string)
|
| 9 |
|
|
ROW FORMAT SERDE 'com.cloudera.science.avro.serde.AvroAsJSONSerde'
|
| 10 |
|
|
INPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerInputFormat'
|
| 11 |
|
|
OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerOutputFormat'
|
| 12 |
|
|
WITH SERDEPROPERTIES
|
| 13 |
|
|
('avro.schema.literal'='{
|
| 14 |
|
|
"namespace": "testing.hive.avro.serde",
|
| 15 |
|
|
"name": "doctors",
|
| 16 |
|
|
"type": "record",
|
| 17 |
|
|
"fields": [
|
| 18 |
|
|
{
|
| 19 |
|
|
"name":"number",
|
| 20 |
|
|
"type":"int",
|
| 21 |
|
|
"doc":"Order of playing the role"
|
| 22 |
|
|
},
|
| 23 |
|
|
{
|
| 24 |
|
|
"name":"first_name",
|
| 25 |
|
|
"type":"string",
|
| 26 |
|
|
"doc":"first name of actor playing role"
|
| 27 |
|
|
},
|
| 28 |
|
|
{
|
| 29 |
|
|
"name":"last_name",
|
| 30 |
|
|
"type":"string",
|
| 31 |
|
|
"doc":"last name of actor playing role"
|
| 32 |
|
|
},
|
| 33 |
|
|
{
|
| 34 |
|
|
"name":"extra_field",
|
| 35 |
|
|
"type":"string",
|
| 36 |
|
|
"doc:":"an extra field not in the original file",
|
| 37 |
|
|
"default":"fishfingers and custard"
|
| 38 |
|
|
}
|
| 39 |
|
|
]
|
| 40 |
|
|
}');
|
| 41 |
|
|
DESCRIBE doctors;
|
| 42 |
|
|
SELECT * from doctors;
|
| 43 |
|
|
SELECT get_json_object(foo, '$.number') from doctors;
|
| 44 |
|
|
|
| 45 |
|
|
Streaming Usage:
|
| 46 |
|
|
|
| 47 |
|
|
hadoop jar hadoop-streaming-2.0.0-mr1-cdh4.1.2.jar \
|
| 48 |
|
|
-libjars avro-json-1.0-SNAPSHOT.jar \
|
| 49 |
|
|
-Dinput.schema.url=file:///doctors.avsc \
|
| 50 |
|
|
-Doutput.schema.url=file:///doctors.avsc \
|
| 51 |
|
|
-inputformat com.cloudera.science.avro.streaming.AvroAsJSONInputFormat \
|
| 52 |
|
|
-outputformat com.cloudera.science.avro.streaming.AvroAsJSONOutputFormat \
|
| 53 |
|
|
-mapper '/bin/cat' \
|
| 54 |
|
|
-input doctors.avro \
|
| 55 |
|
|
-output foo
|
| 56 |
|
|
|
| 57 |
|
|
Streaming Over Multiple Files with Different Schemas (Make sure that the ordering of the
|
| 58 |
|
|
arguments to input.schema.url corresponds to the ordering of the -input args to streaming):
|
| 59 |
|
|
|
| 60 |
|
|
hadoop jar hadoop-streaming-2.0.0-mr1-cdh4.1.2.jar \
|
| 61 |
|
|
-libjars avro-json-1.0-SNAPSHOT.jar \
|
| 62 |
|
|
-Dinput.schema.url=file:///schemas/doctors.avsc,file:///schemas/episodes.avsc \
|
| 63 |
|
|
-inputformat com.cloudera.science.avro.streaming.AvroAsJSONInputFormat \
|
| 64 |
|
|
-mapper '/bin/cat' \
|
| 65 |
|
|
-input /data/doctors.avro \
|
| 66 |
|
|
-input /data/episodes.avro \
|
| 67 |
|
|
-output multijson
|