如何在Spark SQL中指定多个表？

Question

我有一个代码,我需要设置三个表.为此,我需要jdbc为每个表调用三次函数.见下面的代码

val props = new Properties
    props.setProperty("user", "root")
    props.setProperty("password", "pass")

val df0 = sqlContext.read.jdbc(
  "jdbc:mysql://127.0.0.1:3306/Firm42", "company", props)

val df1 = sqlContext.read.jdbc(
  "jdbc:mysql://127.0.0.1:3306/Firm42", "employee", props)

val df2 = sqlContext.read.jdbc(
  "jdbc:mysql://127.0.0.1:3306/Firm42", "company_employee", props)

df0.registerTempTable("company")
df1.registerTempTable("employee")
df2.registerTempTable("company_employee")

val rdf = sqlContext.sql(
  """some_sql_query_with_joins_of_various_tables""".stripMargin)

rdf.show

是否可以简化我的代码？或者也许有一些方法可以在SQL配置中的某处指定多个表.

Answer 1

干:

val url = "jdbc:mysql://127.0.0.1:3306/Firm42"
val tables =  List("company", "employee", "company_employee")

val dfs = for {
    table <- tables
} yield (table, sqlContext.read.jdbc(url, table, props))

for {
  (name, df) <- dfs
} df.registerTempTable(name) 

不需要数据框架？跳过第一个循环:

for {
  table <- tables
} sqlContext.read.jdbc(url, table, props).registerTempTable(table)