[SPARK-24514] Exception while converting RDD to DataFrame

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Invalid
• Affects Version/s: 2.3.0
• Component/s: Java API, Spark Core, SQL

Environment

Spark version 2.3
JDK 1.8
MongoDB Version - 3.4 and 3.6

Sample Data -

{"_id":"5b0d31f892549e10b61d962a","RSEG_MANDT":"800","RSEG_EBELN":"4500017749","RSEG_EBELP":"00020","RSEG_BELNR":"1000000001","RSEG_BUZEI":"000002","RSEG_GJAHR":"2013","RBKP_BUDAT":"2013-10-04","RSEG_MENGE":

{"$numberDecimal":"30.000"}

,"RSEG_LFBNR":"5000000472","RSEG_LFGJA":"2013","RSEG_LFPOS":"0002","NOT_ACCOUNT_MAINTENANCE":

{"$numberDecimal":"1.0000000000"}

,"RBKP_CPUTIMESTAMP":"2013-10-04T10:32:02.000Z","RBKP_WAERS":"USD","RSEG_BNKAN":

{"$numberDecimal":"0.00"}

,"RSEG_WRBTR":

{"$numberDecimal":"2340.00"}

,"RSEG_SHKZG":"S"}

• Flags: Important
• Assignee: Unassigned
• Reporter: SHAILENDRA SHAHANE
• Created: 11/Jun/18 05:57
• Updated: 12/Dec/22 18:10
• Resolved: 12/Jun/18 01:46

Description

I tried to fetch data from MongoDB and got the following exception while converting the RDD to DF.

-----------------Code --------------
SparkSession spark = sparkSessionBuilder.appName("VendorProcessor")
.config("spark.mongodb.input.uri","myMongoUrl")
.config("spark.mongodb.output.uri","myMongoUrl")
.getOrCreate();

JavaSparkContext jsc = new JavaSparkContext(spark.sparkContext());
JavaMongoRDD<Document> rdd = MongoSpark.load(jsc);
Dataset<Row> rbkp = rdd.toDF();
------------------ OR ------------------------

DataFrameReader dfr = spark.read()
.format("com.mongodb.spark.sql")
.option("floatAsBigDecimal", "true");
Dataset<Row> rbkp = dfr.load();

--------------------