Spark Redshift saving into s3 as Parquet - amazon-s3

Question

Issues saving a redshift table into s3 as a parquet file... This is coming from the date field. I'm going to try to convert the column to a long and store it as a unix timestamp for now.
Caused by: java.lang.NumberFormatException: multiple points
at sun.misc.FloatingDecimal.readJavaFormatString(FloatingDecimal.java:1110)
at java.lang.Double.parseDouble(Double.java:540)
at java.text.DigitList.getDouble(DigitList.java:168)
at java.text.DecimalFormat.parse(DecimalFormat.java:1321)
at java.text.SimpleDateFormat.subParse(SimpleDateFormat.java:1793)
at java.text.SimpleDateFormat.parse(SimpleDateFormat.java:1455)
at com.databricks.spark.redshift.Conversions$$anon$1.parse(Conversions.scala:54)
at java.text.DateFormat.parse(DateFormat.java:355)
at com.databricks.spark.redshift.Conversions$.com$databricks$spark$redshift$Conversions$$parseTimestamp(Conversions.scala:67)
at com.databricks.spark.redshift.Conversions$$anonfun$1.apply(Conversions.scala:122)
at com.databricks.spark.redshift.Conversions$$anonfun$1.apply(Conversions.scala:108)
at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244)
at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244)
at scala.collection.IndexedSeqOptimized$class.foreach(IndexedSeqOptimized.scala:33)
at scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:108)
at scala.collection.TraversableLike$class.map(TraversableLike.scala:244)
at scala.collection.mutable.ArrayOps$ofRef.map(ArrayOps.scala:108)
at com.databricks.spark.redshift.Conversions$.com$databricks$spark$redshift$Conversions$$convertRow(Conversions.scala:108)
at com.databricks.spark.redshift.Conversions$$anonfun$createRowConverter$1.apply(Conversions.scala:135)
at com.databricks.spark.redshift.Conversions$$anonfun$createRowConverter$1.apply(Conversions.scala:135)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
at org.apache.spark.sql.execution.datasources.DefaultWriterContainer.writeRows(WriterContainer.scala:241)
... 8 more
These are my gradle dependencies:
dependencies {
compile 'com.amazonaws:aws-java-sdk:1.10.31'
compile 'com.amazonaws:aws-java-sdk-redshift:1.10.31'
compile 'org.apache.spark:spark-core_2.10:1.5.1'
compile 'org.apache.spark:spark-sql_2.10:1.5.1'
compile 'com.databricks:spark-redshift_2.10:0.5.1'
compile 'com.fasterxml.jackson.module:jackson-module-scala_2.10:2.6.3'
}
EDIT 1: df.write.parquet("s3n://bucket/path/log.parquet") is how I'm saving the dataframe after I load in the redshift data using spark-redshift.
EDIT 2: I'm running all of this on my macbook air, maybe too much data corrupts the Dataframe? Not sure... It works when I 'limit 1000', just not for the entire table... So "query" works, but "table" doesn't in the spark-redshift options params.

Answer 1

spark-redshift maintainer here. I believe that the error that you're seeing is caused by a thread-safety bug in spark-redshift (Java DecimalFormat instances are not thread-safe and we were sharing a single instance across multiple threads).
This has been fixed in the 0.5.2 release, which is available on Maven Central and Spark Packages. Upgrade to 0.5.2 and this should work!