2020-04-24 spark json字符串转数组

首先引入包

import org.apache.spark.sql.functions.{explode, from_json, udf}
import org.apache.spark.sql.types._

然后构造数据

val df = spark.sql("""select '[{"item":17991,"score":0.63,"rk":5},{"item":18000,"score":0.64,"rk":4},{"item":15248,"score":0.82,"rk":2},{"item":18008,"score":0.78,"rk":3},{"item":18023,"score":1.0,"rk":1}]' as item""")
df.show(false)

最后使用from_json来构建数组

import spark.implicits._
val jsonDF = df.select(from_json($"item", ArrayType(StructType(StructField("item", StringType) :: StructField("score", DoubleType) ::StructField("rk", IntegerType) :: Nil))).as("items"))
jsonDF.show(false)
jsonDF.printSchema