CODEWITHSUNDEEP
apache-sparkapache-spark-sqlunion-all
Ishan
Ishan

Reputation: 1006

Union of two Spark dataframes with different columns

I am trying to union two Spark dataframes with different set of columns. For this purpose, I referred to following link :-

How to perform union on two DataFrames with different amounts of columns in spark?

My code is as follows - 

val cols1 = finalDF.columns.toSet
val cols2 = df.columns.toSet
val total = cols1 ++ cols2 
finalDF=finalDF.select(expr(cols1, total):_*).unionAll(df.select(expr(cols2, total):_*))

def expr(myCols: Set[String], allCols: Set[String]) = {
  allCols.toList.map(x => x match {
    case x if myCols.contains(x) => col(x)
    case _ => lit(null).as(x)
  })
}

But the problem I am facing is some of the columns in both dataframes are nested. I've columns of both StructType and primitive types. Now, say column A (of StructType) is in df and not in finalDF. But in expr, 

case _ => lit(null).as(x)

is not making it StructType. That's why I am not able to union them. It is giving me following error - 

org.apache.spark.sql.AnalysisException: Union can only be performed on tables with the compatible column types. NullType <> StructType(StructField(_VALUE,StringType,true), StructField(_id,LongType,true)) at the first column of the second table.

Any suggestions what I can do here ?

Upvotes: 1

Views: 14345

Answers (2)

ehacinom
ehacinom

Reputation: 8894

df = df1.join(df2, ['each', 'shared', 'column'], how='full')

will fill missing data with nulls.

Upvotes: 0

Alper t. Turker
Alper t. Turker

Reputation: 35229

I'd use built-in schema inference for this. It is way more expensive, but much simpler than matching complex structures, with possible conflicts:

spark.read.json(df1.toJSON.union(df2.toJSON))

You can also import all files at the same time, and join with information extracted from header, using input_file_name.

import org.apache.spark.sql.function

val metadata: DataFrame  // Just metadata from the header
val data: DataFrame      // All files loaded together

metadata.withColumn("file", input_file_name)
  .join(data.withColumn("file", input_file_name), Seq("file"))

Upvotes: 2

Related Questions