partitionColumn是应该用于确定分区的列。
lowerBound并upperBound确定要获取的值的范围。完整数据集将使用与以下查询对应的行：

SELECT * FROM table WHERE partitionColumn BETWEEN lowerBound AND upperBound
numPartitions确定要创建的分区数。lowerBound和之间的范围upperBound分为numPartitions每个，步幅等于：

upperBound / numPartitions - lowerBound / numPartitions

例如，如果：

lowerBound: 0
upperBound: 1000
numPartitions: 10

步幅等于 100，分区对应于以下查询：

• SELECT * FROM table WHERE partitionColumn BETWEEN 0 AND 100
• SELECT * FROM table WHERE partitionColumn BETWEEN 101 AND 200
• …
• SELECT * FROM table WHERE partitionColumn >= 901

验证

mysql表数据:

验证代码:

import org.apache.spark.TaskContext
import org.apache.spark.sql.SparkSession

object SpkPartApp {

  def main(args: Array[String]): Unit = {

    val spark = SparkSession.builder()
      .master("local[*]")
      .appName(SpkPartApp.getClass.getSimpleName)
      .enableHiveSupport()
      .getOrCreate()

    val jdbcDF = spark.read.format("jdbc")
      .options(Map("url" ->  "jdbc:mysql://127.0.0.1:3306/data_test?user=root&password=123456",
        "dbtable" -> "data_test.test_part",
        "fetchSize" -> "20",
        "partitionColumn" -> "seq",
        "lowerBound" -> "0",
        "upperBound" -> "10",
        "numPartitions" -> "5"))
      .load()

    println(jdbcDF.rdd.getNumPartitions)

    jdbcDF.foreach(row => {
      println("partitionId：" + TaskContext.get.partitionId)
    })
    
    spark.stop()

  }

}