我刚从1.3.1升级到spark 2.0.0,我写了一个简单的代码与hive(1.2.1)交互使用spark sql,我把hive-site.xml放到spark conf目录中,我从sql中获取预期的结果,但它抛出了一个奇怪的AlreadyExistsException(消息:数据库默认已经存在),如何忽略这个？
【码】
from pyspark.sql import SparkSession
ss = SparkSession.builder.appName("test").master("local") \
.config("spark.ui.port", "4041") \
.enableHiveSupport()\
.getOrCreate()
ss.sparkContext.setLogLevel("INFO")
ss.sql("show tables").show()
【登录】
Setting default log level to "WARN".
To adjust logging level use sc.setLogLevel(newLevel).
16/08/08 19:41:22 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/08/08 19:41:24 INFO execution.SparkSqlParser: Parsing command: show tables
16/08/08 19:41:25 INFO hive.HiveUtils: Initializing HiveMetastoreConnection version 1.2.1 using Spark classes.
16/08/08 19:41:26 INFO metastore.HiveMetaStore: 0: Opening raw store with implemenation …