[SPARK-22283] withColumn should replace multiple instances with a single one

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Not A Problem
• Affects Version/s: 2.2.0
• Component/s: Spark Core
• Assignee: Unassigned
• Reporter: Albert Meltzer
• Created: 16/Oct/17 06:32
• Updated: 19/Oct/17 07:12
• Resolved: 19/Oct/17 07:12

Description

Currently, withColumn claims to do the following: "adding a column or replacing the existing column that has the same name."

Unfortunately, if multiple existing columns have the same name (which is a normal occurrence after a join), this results in multiple replaced – and retained –
columns (with the same value), and messages about an ambiguous column.

The current implementation of withColumn contains this:

  def withColumn(colName: String, col: Column): DataFrame = {
    val resolver = sparkSession.sessionState.analyzer.resolver
    val output = queryExecution.analyzed.output
    val shouldReplace = output.exists(f => resolver(f.name, colName))
    if (shouldReplace) {
      val columns = output.map { field =>
        if (resolver(field.name, colName)) {
          col.as(colName)
        } else {
          Column(field)
        }
      }
      select(columns : _*)
    } else {
      select(Column("*"), col.as(colName))
    }
  }

Instead, suggest something like this (which replaces all matching fields with a single instance of the new one):

  def withColumn(colName: String, col: Column): DataFrame = {
    val resolver = sparkSession.sessionState.analyzer.resolver
    val output = queryExecution.analyzed.output
    val existing = output.filterNot(f => resolver(f.name, colName)).map(new Column(_))
    select(existing :+ col.as(colName): _*)
  }