为什么编译器出错但在 Spark-shell 上运行？

Question

The command

println(  sql( "SELECT date_format(now(),'yyyyMMdd')" ).as[String].first )

is working fine on Spark-shell (using Spark v2.2), but the line

 val databaseNow = spark.sql( "SELECT date_format(now(),'yyyyMMdd')" ).as[String].first

not compile. I try many import variations and problem persist. I try syntax variations as df.first.toString but it is not a solution ( [x] is not x ).

Answer 1

Just include the included encoders like:

val spark: SparkSession =
  SparkSession
    .builder()
    .appName("test")
    .getOrCreate()

import spark.implicits._

val databaseNow = spark.sql( "SELECT date_format(now(),'yyyyMMdd')" ).as[String].first