apache-spark - 使用已知架构保存空 DataFrame (Spark 2.2.1)

Question

是否可以使用已知模式保存一个空的 DataFrame 以便将模式写入文件，即使它有 0 条记录？

def example(spark: SparkSession, path: String, schema: StructType) = { 
  val dataframe = spark.createDataFrame(spark.sparkContext.emptyRDD[Row], schema) 
  val dataframeWriter = dataframe.write.mode(SaveMode.Overwrite).format("parquet") 
  dataframeWriter.save(path) 

  spark.read.load(path) // ERROR!! No files to read, so schema unknown 
} 

Answer 1

这是我从 Databricks 支持收到的答案:

This is actually a known issue in Spark. There is already fix done inopensource JIRA -> https://issues.apache.org/jira/browse/SPARK-23271.For more details on how this behavior will change from 2.4 pleasecheck this doc changehttps://github.com/apache/spark/pull/20525/files#diff-d8aa7a37d17a1227cba38c99f9f22511R1808The behavior will be changed from Spark 2.4. Until then you need to gowith any one of the following ways

1. Save a dataframe with at-least one record to preserve its schema
2. Save schema in a JSON file and use later