MongoDB Spark Connector py4j.protocol.Py4JJavaErro

问题:

I have been able to load this MongoDB database before, but am now receiving an error I haven't been able to figure out.

Here is how I start my Spark session:

spark = SparkSession.builder \
        .master("local[*]") \
        .appName("collab_rec") \
        .config("spark.mongodb.input.uri", "mongodb://127.0.0.1/example.collection") \
        .config("spark.mongodb.output.uri", "mongodb://127.0.0.1/example.collection") \
        .getOrCreate()

I run this script so that I can interact with spark through ipython wich loads the mongo spark connector package:

#!/bin/bash
export PYSPARK_DRIVER_PYTHON=ipython

${SPARK_HOME}/bin/pyspark \
--master local[4] \
--executor-memory 1G \
--driver-memory 1G \
--conf spark.sql.warehouse.dir="file:///tmp/spark-warehouse" \
--packages com.databricks:spark-csv_2.11:1.5.0 \
--packages com.amazonaws:aws-java-sdk-pom:1.10.34 \
--packages org.apache.hadoop:hadoop-aws:2.7.3 \
--packages org.mongodb.spark:mongo-spark-connector_2.11:2.0.0\

Spark loads fine and it appears the package is loading correctly as well.

Here is how I attempt to load that database into a dataframe:

df = spark.read.format("com.mongodb.spark.sql.DefaultSource").load()

However, on that line, I receive the following error:

Py4JJavaError: An error occurred while calling o46.load.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 0.0 failed 1 times, most recent failure: Lost task 0.0 in stage 0.0 (TID 0, localhost, executor driver): java.lang.NoSuchMethodError: org.apache.spark.sql.catalyst.analysis.TypeCoercion$.findTightestCommonTypeOfTwo()Lscala/Function2;
    at com.mongodb.spark.sql.MongoInferSchema$.com$mongodb$spark$sql$MongoInferSchema$$compatibleType(MongoInferSchema.scala:132)
    at com.mongodb.spark.sql.MongoInferSchema$$anonfun$3.apply(MongoInferSchema.scala:76)
    at com.mongodb.spark.sql.MongoInferSchema$$anonfun$3.apply(MongoInferSchema.scala:76)

From what I can see through the following documentation/tutorial I am attempting to load the dataframe correctly:

https://docs.mongodb.com/spark-connector/master/python-api/

I am using Spark 2.2.0 Note that I have been able to replicate this error on both my mac and linux through AWS.

回答1:

I figured out the answer to my question. This was a compatibility issue with the Mongo-Spark connector and the version of Spark that I upgraded to. Specifically, the findTightestCommonTypeOfTwo value was renamed in the PR:

https://github.com/apache/spark/pull/16786/files

For Spark 2.2.0 the compatible Mongo-Spark connector is also 2.2.0, thus in my example, the package would be loaded like this:

--packages org.mongodb.spark:mongo-spark-connector_2.11:2.2.0\

This could change in the future so when using the connector, you should check for compatibility with the version of Spark being used.