所用的软件版本:
spark2.3.0
hbase1.4.6
IDEA2019.1
在利用spark.read.json("e:/test.json")读取多行的json文件,利用DataFrame的DF.show()显示,除了_corrupt_record这行有全部的json内容,其他column全部都是null,DF.printSchema()的结果是:
root
|-- _corrupt_record: string (nullable = true)
https://stackoverflow.com/questions/38545850/read-multiline-json-in-apache-spark
https://stackoverflow.com/questions/49427623/could-not-read-data-from-json-using-pyspark
https://stackoverflow.com/questions/38895057/reading-json-with-apache-spark-corrupt-record
网上查询发现要添加下面这一行:
spark.read
.option("multiLine", true).option("mode", "PERMISSIVE")
.json("e:/test.json")
做好记录,供以后参考!!