Spark Connector Python Guide¶

On this page

Prerequisites
Getting Started
- Python Spark Shell
- Create a SparkSession Object
Tutorials

Source Code

For the source code that contains the examples below, see introduction.py.

Prerequisites¶

Basic working knowledge of MongoDB and Apache Spark. Refer to the MongoDB documentation and Spark documentation for more details.
Running MongoDB instance (version 2.6 or later).
Spark 2.0.x.
Scala 2.11.x

Getting Started¶

Python Spark Shell¶

This tutorial uses the pyspark shell, but the code works with self-contained Python applications as well.

When starting the pyspark shell, you can specify:

the --packages option to download the MongoDB Spark Connector package. The following package is available:
- mongo-spark-connector_2.11 for use with Scala 2.11.x
the --conf option to configure the MongoDB Spark Connnector. These settings configure the SparkConf object.

Note

When specifying the Connector configuration via SparkConf, you must prefix the settings appropriately. For details and other available MongoDB Spark Connector options, see the Configuration Options.

The following example starts the pyspark shell from the command line:

copy

./bin/pyspark --conf "spark.mongodb.input.uri=mongodb://127.0.0.1/test.myCollection?readPreference=primaryPreferred" \
              --conf "spark.mongodb.output.uri=mongodb://127.0.0.1/test.myCollection" \
              --packages org.mongodb.spark:mongo-spark-connector_2.11:2.0.0

The spark.mongodb.input.uri specifies the MongoDB server address (127.0.0.1), the database to connect (test), and the collection (myCollection) from which to read data, and the read preference.
The spark.mongodb.output.uri specifies the MongoDB server address (127.0.0.1), the database to connect (test), and the collection (myCollection) to which to write data. Connects to port 27017 by default.
The packages option specifies the Spark Connector’s Maven coordinates, in the format groupId:artifactId:version.

The examples in this tutorial will use this database and collection.

Create a `SparkSession` Object¶

Note

When you start pyspark you get a SparkSession object called spark by default. In a standalone Python application, you need to create your SparkSession object explicitly, as show below.

If you specified the spark.mongodb.input.uri and spark.mongodb.output.uri configuration options when you started pyspark, the default SparkSession object uses them. If you’d rather create your own SparkSession object from within pyspark, you can use SparkSession.builder and specify different configuration options.

copy

from pyspark.sql import SparkSession

my_spark = SparkSession \
    .builder \
    .appName("myApp") \
    .config("spark.mongodb.input.uri", "mongodb://127.0.0.1/test.coll") \
    .config("spark.mongodb.output.uri", "mongodb://127.0.0.1/test.coll") \
    .getOrCreate()

You can use a SparkSession object to write data to MongoDB, read data from MongoDB, create DataFrames, and perform SQL operations.

Tutorials¶

← Datasets and SQL Write to MongoDB →