java.lang.NullPointerException when save as TFRecords

[SoloBean]

Caused by: java.lang.NullPointerException
at org.tensorflow.spark.datasources.tfrecords.serde.DefaultTfRecordRowEncoder$.$anonfun$encodeFeature$5(DefaultTfRecordRowEncoder.scala:136)
at scala.collection.TraversableLike.$anonfun$map$1(TraversableLike.scala:238)
at scala.collection.IndexedSeqOptimized.foreach(IndexedSeqOptimized.scala:36)
at scala.collection.IndexedSeqOptimized.foreach$(IndexedSeqOptimized.scala:33)
at scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:198)
at scala.collection.TraversableLike.map(TraversableLike.scala:238)
at scala.collection.TraversableLike.map$(TraversableLike.scala:231)
at scala.collection.mutable.ArrayOps$ofRef.map(ArrayOps.scala:198)
at org.tensorflow.spark.datasources.tfrecords.serde.DefaultTfRecordRowEncoder$.encodeFeature(DefaultTfRecordRowEncoder.scala:136)
at org.tensorflow.spark.datasources.tfrecords.serde.DefaultTfRecordRowEncoder$.$anonfun$encodeExample$1(DefaultTfRecordRowEncoder.scala:64)
at scala.collection.immutable.List.foreach(List.scala:392)
at org.tensorflow.spark.datasources.tfrecords.serde.DefaultTfRecordRowEncoder$.encodeExample(DefaultTfRecordRowEncoder.scala:61)
at org.tensorflow.spark.datasources.tfrecords.DefaultSource.$anonfun$createRelation$3(DefaultSource.scala:61)
at scala.collection.Iterator$$anon$10.next(Iterator.scala:459)
at org.apache.spark.internal.io.SparkHadoopWriter$.$anonfun$executeTask$1(SparkHadoopWriter.scala:132)
at org.apache.spark.util.Utils$.tryWithSafeFinallyAndFailureCallbacks(Utils.scala:1452)
at org.apache.spark.internal.io.SparkHadoopWriter$.executeTask(SparkHadoopWriter.scala:130)
... 9 more

what is data like can case this exception?