An example of bidirectional communication between a MongoDB collection and Apache Hive.

Hive-Mongo

# Git clone, build, and copy the necessary MongoDB jars to your worker node $HADOOP_HOME/lib directories
cd ~/
git clone https://github.com/mongodb/mongo-hadoop
cd mongo-hadoop
./gradlew jar 
sudo cp build/libs/* /usr/lib/hadoop/lib
sudo cp core/build/libs/* /usr/lib/hadoop/lib
sudo cp hive/build/libs/* /usr/lib/hadoop/lib

cd ~/
git clone https://github.com/mongodb/mongo-java-driver
cd mongo-java-driver
./gradlew jar 
sudo cp build/libs/* /usr/lib/hadoop/lib

hive> create table mongo_table(example_column string)
> stored by 'com.mongodb.hadoop.hive.MongoStorageHandler'
> tblproperties('mongo.uri'='mongodb://localhost:27017/test.test');

hive> insert into table mongo_table select * from hive_table;

hive> add jar /usr/lib/hadoop/lib/mongo-hadoop-hive-*.jar;
hive> add jar /usr/lib/hadoop/lib/mongo-java-driver.jar;
hive> create table mongo_mapped_table(hive_column string)
> stored by 'com.mongodb.hadoop.hive.MongoStorageHandler'
> with serdeproperties ('mongo.columns.mapping'='{"hive_column":"mongo_field"}')
> tblproperties('mongo.uri'='mongodb://n0.dev:27017/test.test');

#If using an authSource param, mongo.user and mongo.passwd as table properties will not work. Instead you'll need to specify user/pw in the URI, as in: 'mongo.uri'='mongodb://user:password@localhost/test.test?authSource=admin'

hive> insert into table mongo_mapped_table select * from hive_table;

hive> select * from mongo_table;
hive> select * from mongo_mapped_table;

Known to work with HDP 2.1.4, hive.execution.engine may be mr OR tez.